diff --git a/haystack/components/converters/xlsx.py b/haystack/components/converters/xlsx.py
index 370e1615f7..5b0aff95bb 100644
--- a/haystack/components/converters/xlsx.py
+++ b/haystack/components/converters/xlsx.py
@@ -227,10 +227,15 @@ def _extract_tables(self, bytestream: ByteStream) -> tuple[list[str], list[dict]
                     "index": True,
                     "headers": value.columns,
                     "tablefmt": "pipe",
+                    "missingval": "",
                     **self.table_format_kwargs,
                 }
-                # to_markdown uses tabulate
-                tables.append(value.to_markdown(**resolved_kwargs))
+                # to_markdown uses tabulate, whose missingval only covers None: a NaN
+                # reaches the formatter as a number and is written out as "nan". Replace
+                # the empty cells with None so an empty cell reads as empty, the way
+                # to_csv already writes it, and so missingval keeps working.
+                filled = value.astype(object).where(value.notna(), None)
+                tables.append(filled.to_markdown(**resolved_kwargs))
             # add sheet_name to metadata
             metadata.append({"xlsx": {"sheet_name": key}})
         return tables, metadata
diff --git a/releasenotes/notes/xlsx-markdown-missing-value-3f0c1de84b6a29c7.yaml b/releasenotes/notes/xlsx-markdown-missing-value-3f0c1de84b6a29c7.yaml
new file mode 100644
index 0000000000..fbf5b9629b
--- /dev/null
+++ b/releasenotes/notes/xlsx-markdown-missing-value-3f0c1de84b6a29c7.yaml
@@ -0,0 +1,8 @@
+---
+fixes:
+  - |
+    Fixed `XLSXToDocument` writing the string `nan` into an empty cell when
+    `table_format="markdown"`. The same cell is written as an empty field by
+    `table_format="csv"`, so an empty cell now reads as empty in both formats.
+    A different placeholder can be set with
+    `table_format_kwargs={"missingval": "N/A"}`.
diff --git a/test/components/converters/test_xlsx_to_document.py b/test/components/converters/test_xlsx_to_document.py
index 3e75a77a52..a55623aeaf 100644
--- a/test/components/converters/test_xlsx_to_document.py
+++ b/test/components/converters/test_xlsx_to_document.py
@@ -86,7 +86,8 @@ def test_run_markdown(self, test_files_path: Path) -> None:
         }
         assert (
             documents[1].content
-            == "|    | A     | B     |\n|---:|:------|:------|\n|  1 | col_c | col_d |\n|  2 | True  | nan   |"
+            # The empty cell reads as empty, the way the CSV format already writes it.
+            == "|    | A     | B     |\n|---:|:------|:------|\n|  1 | col_c | col_d |\n|  2 | True  |       |"
         )
         assert documents[1].meta == {
             "date_added": "2022-01-01T00:00:00",
@@ -94,6 +95,16 @@ def test_run_markdown(self, test_files_path: Path) -> None:
             "xlsx": {"sheet_name": "Table Missing Value"},
         }
 
+    def test_run_markdown_missing_value(self, test_files_path: Path) -> None:
+        """table_format_kwargs["missingval"] reaches tabulate for an empty cell."""
+        converter = XLSXToDocument(table_format="markdown", table_format_kwargs={"missingval": "N/A"})
+        paths: list[str | Path | ByteStream] = [test_files_path / "xlsx" / "basic_tables_two_sheets.xlsx"]
+        results = converter.run(sources=paths)
+        assert (
+            results["documents"][1].content
+            == "|    | A     | B     |\n|---:|:------|:------|\n|  1 | col_c | col_d |\n|  2 | True  | N/A   |"
+        )
+
     @pytest.mark.parametrize(
         "sheet_name, expected_sheet_name, expected_content",
         [