diff --git a/test/test_um2netcdf.py b/test/test_um2netcdf.py
index 92bb268..a88811c 100644
--- a/test/test_um2netcdf.py
+++ b/test/test_um2netcdf.py
@@ -15,6 +15,8 @@
 import iris.coords
 import iris.exceptions
 
+from netCDF4 import default_fillvals
+
 D_LAT_N96 = 1.25   # Degrees between latitude points on N96 grid
 D_LON_N96 = 1.875  # Degrees between longitude points on N96 grid
 
@@ -1207,3 +1209,56 @@ def test_convert_32_bit_with_float64(ua_plev_cube):
     ua_plev_cube.data = array
     um2nc.convert_32_bit(ua_plev_cube)
     assert ua_plev_cube.data.dtype == np.float32
+
+
+@pytest.mark.parametrize(
+    "cube_data, custom_fill_val, expected_fill_val",
+    [
+        (np.array([1.1, 2.1], dtype="float32"),
+         None,
+         np.float32(um2nc.DEFAULT_FILL_VAL_FLOAT)),
+        (np.array([1.1, 2.1], dtype="float64"),
+         None,
+         np.float64(um2nc.DEFAULT_FILL_VAL_FLOAT)),
+        (np.array([1.1, 2.1], dtype="complex64"),
+         None,
+         np.complex64(default_fillvals["c8"])),
+        (np.array([1, 2], dtype="int32"),
+         None,
+         np.int32(default_fillvals["i4"])),
+        (np.array([1, 2], dtype=np.dtype("int64")),
+         None,
+         np.int64(default_fillvals["i8"])),
+        (np.array([1, 2], dtype=np.dtype("int64")),
+         -12345,
+         np.int64(-12345))
+    ]
+)
+def test_fix_fill_value(cube_data, custom_fill_val, expected_fill_val):
+    """
+    Check that correct default and custom fill values are added based
+    on a cube's data's type.
+    """
+    fake_cube = DummyCube(12345, "fake_var", attributes={})
+    fake_cube.data = cube_data
+
+    um2nc.fix_fill_value(fake_cube, custom_fill_val)
+
+    cube_fill_val = fake_cube.attributes["missing_value"]
+
+    assert cube_fill_val[0] == expected_fill_val
+    # Check new fill value type matches cube's data's type
+    assert cube_fill_val.dtype == cube_data.dtype
+
+
+def test_fix_fill_value_wrong_type():
+    """
+    Check that an error is raised when a custom fill value's
+    type does not match the cube's data's type.
+    """
+    fake_cube = DummyCube(12345, "fake_var", attributes={})
+    fake_cube.data = np.array([1, 2, 3], dtype="int32")
+    custom_fill_val = np.float32(151.11)
+
+    with pytest.raises(TypeError):
+        um2nc.fix_fill_value(fake_cube, custom_fill_val)
diff --git a/umpost/um2netcdf.py b/umpost/um2netcdf.py
index de886d3..b0f52d6 100644
--- a/umpost/um2netcdf.py
+++ b/umpost/um2netcdf.py
@@ -70,6 +70,8 @@
 LEVEL_HEIGHT = "level_height"
 SIGMA = "sigma"
 
+DEFAULT_FILL_VAL_FLOAT = 1.e20
+
 
 class PostProcessingError(Exception):
     """Generic class for um2nc specific errors."""
@@ -319,6 +321,35 @@ def fix_latlon_coords(cube, grid_type, dlat, dlon):
     fix_lon_coord_name(longitude_coordinate, grid_type, dlon)
 
 
+def fix_fill_value(cube, custom_fill_val=None):
+    """
+    Set a cube's missing_value attribute according to the data's dtype.
+
+    Parameters
+    ----------
+    cube: Iris cube object (modified in place).
+    custom_fill_val: Optional custom fill value. Type should match
+        the cube data's type.
+    """
+    if custom_fill_val is not None:
+        if type(custom_fill_val) == cube.data.dtype:
+            fill_value = custom_fill_val
+        else:
+            msg = (f"custom_fill_val type {type(custom_fill_val)} does not "
+                   f"match cube {cube.name()} data type {cube.data.dtype}.")
+            raise TypeError(msg)
+
+    elif cube.data.dtype.kind == 'f':
+        fill_value = DEFAULT_FILL_VAL_FLOAT
+    else:
+        fill_value = default_fillvals[
+            f"{cube.data.dtype.kind:s}{cube.data.dtype.itemsize:1d}"
+        ]
+
+    # Use an array to force the type to match the data type
+    cube.attributes['missing_value'] = np.array([fill_value], cube.data.dtype)
+
+
 # TODO: split cube ops into functions, this will likely increase process() workflow steps
 def cubewrite(cube, sman, compression, use64bit, verbose):
     # TODO: move into process() AND if a new cube is returned, swap into filtered cube list
@@ -328,15 +359,7 @@ def cubewrite(cube, sman, compression, use64bit, verbose):
     if not use64bit:
         convert_32_bit(cube)
 
-    # Set the missing_value attribute. Use an array to force the type to match
-    # the data type
-    if cube.data.dtype.kind == 'f':
-        fill_value = 1.e20
-    else:
-        # Use netCDF defaults
-        fill_value = default_fillvals['%s%1d' % (cube.data.dtype.kind, cube.data.dtype.itemsize)]
-
-    cube.attributes['missing_value'] = np.array([fill_value], cube.data.dtype)
+    fix_fill_value(cube)
 
     # If reference date is before 1600 use proleptic gregorian
     # calendar and change units from hours to days