diff --git a/asdf/_tests/tags/core/tests/test_ndarray.py b/asdf/_tests/tags/core/tests/test_ndarray.py
index 5688a3681..02c79f469 100644
--- a/asdf/_tests/tags/core/tests/test_ndarray.py
+++ b/asdf/_tests/tags/core/tests/test_ndarray.py
@@ -812,6 +812,7 @@ def test_structured_datatype_validation(tmp_path):
        data: [[1, 'a'], [2, 'b'], [3, 'c']]
        datatype:
          - name: a
+           description: a description
            datatype: int8
          - name: b
            datatype: ['ascii', 8]
@@ -829,6 +830,7 @@ def test_structured_datatype_validation(tmp_path):
          - name: a
            datatype: int64
          - name: b
+           title: a title
            datatype: ['ascii', 8]
     """
     buff = helpers.yaml_to_asdf(content)
diff --git a/asdf/tags/core/ndarray.py b/asdf/tags/core/ndarray.py
index 37574073f..df78e771f 100644
--- a/asdf/tags/core/ndarray.py
+++ b/asdf/tags/core/ndarray.py
@@ -7,6 +7,8 @@
 from asdf import util
 from asdf._jsonschema import ValidationError
 
+_STRUCTURED_DATATYPE_KEYS = {"name", "datatype", "byteorder", "shape"}
+
 _datatype_names = {
     "int8": "i1",
     "int16": "i2",
@@ -533,6 +535,13 @@ def validate_datatype(validator, datatype, instance, schema):
         msg = "Not an array"
         raise ValidationError(msg)
 
+    # We are only concerned with some fields from the datatype
+    # object in the schema so if the schema datatype is structured
+    # copy the datatype and drop the irrelevant fields
+    # name datatype byteorder shape
+    if isinstance(datatype, list) and len(datatype) and isinstance(datatype[0], dict):
+        datatype = [{k: v for k, v in subitem.items() if k in _STRUCTURED_DATATYPE_KEYS} for subitem in datatype]
+
     if datatype == in_datatype:
         return
 
diff --git a/changes/1901.bugfix.rst b/changes/1901.bugfix.rst
new file mode 100644
index 000000000..b3066d871
--- /dev/null
+++ b/changes/1901.bugfix.rst
@@ -0,0 +1 @@
+Allow extra keywords in structured datatype validation.