diff --git a/sdm/processing/__init__.py b/sdm/processing/__init__.py
index 45a6d4a48..800b423ab 100644
--- a/sdm/processing/__init__.py
+++ b/sdm/processing/__init__.py
@@ -35,6 +35,7 @@
     RankGaussian,
     Standardize,
     RobustScale,
+    MinMaxScale,
     FlipSign,
     DropConstantColumns,
     PCA,
@@ -83,6 +84,7 @@
     "RankGaussian",
     "Standardize",
     "RobustScale",
+    "MinMaxScale",
     "FlipSign",
     "DropConstantColumns",
     "PCA",
diff --git a/sdm/processing/numerical/__init__.py b/sdm/processing/numerical/__init__.py
index d8665649d..673cfff32 100644
--- a/sdm/processing/numerical/__init__.py
+++ b/sdm/processing/numerical/__init__.py
@@ -13,6 +13,7 @@
 from sdm.processing.numerical.rank_gaussian import RankGaussian
 from sdm.processing.numerical.standardize import Standardize
 from sdm.processing.numerical.robust_scale import RobustScale
+from sdm.processing.numerical.min_max_scale import MinMaxScale
 from sdm.processing.numerical.flip_sign import FlipSign
 from sdm.processing.numerical.constant import DropConstantColumns
 from sdm.processing.numerical.pca import PCA
@@ -32,6 +33,7 @@
     "RankGaussian",
     "Standardize",
     "RobustScale",
+    "MinMaxScale",
     "FlipSign",
     "DropConstantColumns",
     "PCA",
diff --git a/sdm/processing/numerical/min_max_scale.py b/sdm/processing/numerical/min_max_scale.py
new file mode 100644
index 000000000..40a3ab6da
--- /dev/null
+++ b/sdm/processing/numerical/min_max_scale.py
@@ -0,0 +1,57 @@
+# SPDX-FileCopyrightText: Copyright (c) 2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+# SPDX-License-Identifier: Apache-2.0
+
+import math
+
+import torch
+from torch import Tensor
+
+from sdm import Stype, TableTensor
+from sdm.processing import InvertibleMixin, Processor
+from sdm.processing.numerical._stats import _isfinite
+
+
+class MinMaxScale(Processor, InvertibleMixin):
+    """Scale each feature column by its minimum and range.
+
+    Values within the fitted range map to ``[0, 1]``; values outside it are
+    not clipped. Constant columns use a unit scale to keep the transform
+    finite and invertible. NaN and infinite values are ignored when fitting
+    statistics and preserved during the transform. Columns without finite
+    fitted values transform to NaN.
+    """
+
+    handles_stypes = (Stype.numerical,)
+    requires_fit = True
+
+    minimum: Tensor
+    scale: Tensor
+
+    def __init__(self) -> None:
+        super().__init__()
+        self.register_buffer("minimum", torch.empty(0))
+        self.register_buffer("scale", torch.empty(0))
+
+    def _fit(
+        self,
+        table: TableTensor,
+        *,
+        generator: torch.Generator | None = None,
+    ) -> None:
+        numerical = table.numerical
+        nonfinite = ~_isfinite(numerical)
+        lower = numerical.masked_fill(nonfinite, math.inf)
+        lower = lower.amin(dim=-2, keepdim=True)
+        upper = numerical.masked_fill(nonfinite, -math.inf)
+        upper = upper.amax(dim=-2, keepdim=True)
+        self._set_buffer("minimum", lower)
+        scale = torch.where(lower == upper, 1.0, upper - lower)
+        self._set_buffer("scale", scale)
+
+    def _transform(self, table: TableTensor) -> TableTensor:
+        numerical = (table.numerical - self.minimum).div_(self.scale)
+        return table.replace_blocks(numerical=numerical)
+
+    def _inverse_transform(self, table: TableTensor) -> TableTensor:
+        numerical = (table.numerical * self.scale).add_(self.minimum)
+        return table.replace_blocks(numerical=numerical)
diff --git a/test/processing/numerical/test_min_max_scale.py b/test/processing/numerical/test_min_max_scale.py
new file mode 100644
index 000000000..1621c6869
--- /dev/null
+++ b/test/processing/numerical/test_min_max_scale.py
@@ -0,0 +1,128 @@
+# SPDX-FileCopyrightText: Copyright (c) 2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+# SPDX-License-Identifier: Apache-2.0
+
+import torch
+
+from sdm import TableTensor
+from sdm.processing import MinMaxScale
+from sdm.testing import withCUDA
+
+
+@withCUDA
+def test_min_max_scale_maps_fitted_range_to_unit_interval(
+    device: torch.device,
+) -> None:
+    inp = torch.tensor(
+        [[1.0, 5.0], [3.0, 5.0], [5.0, 5.0]],
+        dtype=torch.float64,
+        device=device,
+    )
+    table = TableTensor.from_tensor(inp)
+
+    processor = MinMaxScale().fit(table)
+    out = processor.transform(table)
+
+    expected = torch.tensor(
+        [[0.0, 0.0], [0.5, 0.0], [1.0, 0.0]],
+        dtype=torch.float64,
+        device=device,
+    )
+    torch.testing.assert_close(out.numerical, expected)
+
+
+@withCUDA
+def test_min_max_scale_does_not_clip_values_outside_the_fitted_table(
+    device: torch.device,
+) -> None:
+    context = torch.tensor(
+        [[1.0, 5.0], [3.0, 5.0]],
+        dtype=torch.float64,
+        device=device,
+    )
+    query = torch.tensor(
+        [[-1.0, 3.0], [7.0, 9.0]],
+        dtype=torch.float64,
+        device=device,
+    )
+
+    processor = MinMaxScale().fit(TableTensor.from_tensor(context))
+    out = processor.transform(TableTensor.from_tensor(query))
+
+    expected = torch.tensor(
+        [[-1.0, -2.0], [3.0, 4.0]],
+        dtype=torch.float64,
+        device=device,
+    )
+    torch.testing.assert_close(out.numerical, expected)
+
+
+@withCUDA
+def test_min_max_scale_preserves_nan_and_inf(device: torch.device) -> None:
+    inp = torch.tensor(
+        [
+            [1.0, 5.0],
+            [2.0, float("nan")],
+            [3.0, 5.0],
+            [float("inf"), 5.0],
+            [-float("inf"), float("inf")],
+        ],
+        dtype=torch.float64,
+        device=device,
+    )
+
+    out = MinMaxScale().fit_transform(TableTensor.from_tensor(inp))
+
+    expected = torch.tensor(
+        [
+            [0.0, 0.0],
+            [0.5, float("nan")],
+            [1.0, 0.0],
+            [float("inf"), 0.0],
+            [-float("inf"), float("inf")],
+        ],
+        dtype=torch.float64,
+        device=device,
+    )
+    torch.testing.assert_close(out.numerical, expected, equal_nan=True)
+
+
+@withCUDA
+def test_min_max_scale_fits_leading_batches_independently(
+    device: torch.device,
+) -> None:
+    context = torch.tensor(
+        [[[1.0], [3.0]], [[10.0], [14.0]]],
+        dtype=torch.float64,
+        device=device,
+    )
+    query = torch.tensor(
+        [[[2.0]], [[12.0]]],
+        dtype=torch.float64,
+        device=device,
+    )
+
+    processor = MinMaxScale().fit(TableTensor.from_tensor(context))
+    out = processor.transform(TableTensor.from_tensor(query))
+
+    expected = torch.full((2, 1, 1), 0.5, dtype=torch.float64, device=device)
+    torch.testing.assert_close(out.numerical, expected)
+
+
+@withCUDA
+def test_min_max_scale_all_nonfinite_fit_yields_nan(
+    device: torch.device,
+) -> None:
+    fit = torch.tensor(
+        [[float("nan")], [float("inf")], [-float("inf")]],
+        dtype=torch.float64,
+        device=device,
+    )
+    query = torch.tensor([[1.0], [2.0]], dtype=torch.float64, device=device)
+
+    out = (
+        MinMaxScale()
+        .fit(TableTensor.from_tensor(fit))
+        .transform(TableTensor.from_tensor(query))
+    )
+
+    assert out.numerical.isnan().all()
diff --git a/test/processing/test_contract.py b/test/processing/test_contract.py
index 7525b3b8c..8078d86d6 100644
--- a/test/processing/test_contract.py
+++ b/test/processing/test_contract.py
@@ -129,6 +129,7 @@ def _make_processor_pair(
     ProcessorCase(sp.Standardize()),
     ProcessorCase(sp.RankGaussian()),
     ProcessorCase(sp.RobustScale()),
+    ProcessorCase(sp.MinMaxScale()),
     ProcessorCase(sp.FlipSign()),
     ProcessorCase(sp.DropConstantColumns()),
     ProcessorCase(sp.PCA(2)),