diff --git a/src/database/datasets.py b/src/database/datasets.py
index 4e76dcf9..26eb33d8 100644
--- a/src/database/datasets.py
+++ b/src/database/datasets.py
@@ -1,6 +1,7 @@
 """Translation from https://github.com/openml/OpenML/blob/c19c9b99568c0fabb001e639ff6724b9a754bbc9/openml_OS/models/api/v1/Api_data.php#L707."""
 
 import datetime
+from collections import defaultdict
 
 from sqlalchemy import text
 from sqlalchemy.engine import Row
@@ -134,6 +135,26 @@ async def get_features(dataset_id: int, connection: AsyncConnection) -> list[Fea
     return [Feature(**row, nominal_values=None) for row in rows]
 
 
+async def get_feature_ontologies(
+    dataset_id: int,
+    connection: AsyncConnection,
+) -> dict[int, list[str]]:
+    rows = await connection.execute(
+        text(
+            """
+            SELECT `index`, `value`
+            FROM data_feature_description
+            WHERE `did` = :dataset_id AND `description_type` = 'ontology'
+            """,
+        ),
+        parameters={"dataset_id": dataset_id},
+    )
+    ontologies: dict[int, list[str]] = defaultdict(list)
+    for row in rows.mappings():
+        ontologies[row["index"]].append(row["value"])
+    return ontologies
+
+
 async def get_feature_values(
     dataset_id: int,
     *,
diff --git a/src/routers/openml/datasets.py b/src/routers/openml/datasets.py
index d86ed848..164efd7d 100644
--- a/src/routers/openml/datasets.py
+++ b/src/routers/openml/datasets.py
@@ -294,6 +294,10 @@ async def get_dataset_features(
     assert expdb is not None  # noqa: S101
     await _get_dataset_raise_otherwise(dataset_id, user, expdb)
     features = await database.datasets.get_features(dataset_id, expdb)
+    ontologies = await database.datasets.get_feature_ontologies(dataset_id, expdb)
+    for feature in features:
+        feature.ontology = ontologies.get(feature.index)
+
     for feature in [f for f in features if f.data_type == FeatureType.NOMINAL]:
         feature.nominal_values = await database.datasets.get_feature_values(
             dataset_id,
diff --git a/src/schemas/datasets/openml.py b/src/schemas/datasets/openml.py
index b1c23a73..767cbb81 100644
--- a/src/schemas/datasets/openml.py
+++ b/src/schemas/datasets/openml.py
@@ -40,6 +40,7 @@ class Feature(BaseModel):
     index: int
     name: str
     data_type: FeatureType
+    ontology: list[str] | None = None
     is_target: bool
     is_ignore: bool
     is_row_identifier: bool
diff --git a/tests/routers/openml/migration/datasets_migration_test.py b/tests/routers/openml/migration/datasets_migration_test.py
index 5ff6fe86..75f30863 100644
--- a/tests/routers/openml/migration/datasets_migration_test.py
+++ b/tests/routers/openml/migration/datasets_migration_test.py
@@ -259,6 +259,8 @@ async def test_datasets_feature_is_identical(
                 values = feature.pop(key)
                 # The old API returns a str if there is only a single element
                 feature["nominal_value"] = values if len(values) > 1 else values[0]
+            elif key == "ontology":
+                del feature[key]  # Added back in with follow up PR #262
             else:
                 # The old API formats bool as string in lower-case
                 feature[key] = str(value) if not isinstance(value, bool) else str(value).lower()