diff --git a/CHANGELOG.md b/CHANGELOG.md
index dd593b038..b2a5374ca 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -34,6 +34,7 @@ This change was made for consistency between row-height governed by text or imag
 * Edge case when parsing a Markdown link followed by a newline - _cf._ issue [#916](https://github.com/py-pdf/fpdf2/issues/916), and when bold/italics/underline markers are repeated
 * Zoom not set correctly when a numeric value was set in [`set_display_mode()`](https://py-pdf.github.io/fpdf2/fpdf/fpdf.html#fpdf.fpdf.FPDF.set_display_mode) - _cf._ issue [#926](https://github.com/py-pdf/fpdf2/issues/926)
 * [`FPDF.table()`](https://py-pdf.github.io/fpdf2/fpdf/fpdf.html#fpdf.fpdf.FPDF.table): images no longer overlap with cell borders - _cf._ issue [#892](https://github.com/py-pdf/fpdf2/issues/892)
+* Encryption of strings containing non-latin characters - __cf.__ issue [#933](https://github.com/py-pdf/fpdf2/issues/933)
 
 ## [2.7.5] - 2023-08-04
 ### Added
diff --git a/fpdf/encryption.py b/fpdf/encryption.py
index f4b872cb9..c2af31ac8 100644
--- a/fpdf/encryption.py
+++ b/fpdf/encryption.py
@@ -11,6 +11,8 @@
 import math
 import stringprep
 import unicodedata
+from binascii import hexlify
+from codecs import BOM_UTF16_BE
 from os import urandom
 from typing import Callable, Iterable, Type, Union
 
@@ -220,7 +222,11 @@ def encrypt_string(self, string: str, obj_id: int) -> str:
         if self.encryption_method == EncryptionMethod.NO_ENCRYPTION:
             return PDFString(string, encrypt=False).serialize()
         LOGGER.debug("Encrypting string: %s", string)
-        return f"<{bytes(self.encrypt_bytes(string.encode('latin-1'), obj_id)).hex().upper()}>"
+        try:
+            string.encode("latin-1")
+            return f"<{bytes(self.encrypt_bytes(string.encode('latin-1'), obj_id)).hex().upper()}>"
+        except UnicodeEncodeError:
+            return f'<{hexlify(bytearray(self.encrypt_bytes(BOM_UTF16_BE + string.encode("utf-16-be"), obj_id))).decode("latin-1")}>'
 
     def encrypt_stream(self, stream: bytes, obj_id: int) -> bytes:
         if self.encryption_method == EncryptionMethod.NO_ENCRYPTION:
diff --git a/test/encryption/encryption_unicode.pdf b/test/encryption/encryption_unicode.pdf
new file mode 100644
index 000000000..a260d4cf9
Binary files /dev/null and b/test/encryption/encryption_unicode.pdf differ
diff --git a/test/encryption/test_encryption.py b/test/encryption/test_encryption.py
index d41fd8389..18b1bce47 100644
--- a/test/encryption/test_encryption.py
+++ b/test/encryption/test_encryption.py
@@ -293,3 +293,20 @@ def test_password_prep():
     with pytest.raises(FPDFException) as e:
         sh.prepare_string("\u0627\x31")  # Error - bidirectional check
     assert sh.prepare_string("A" * 300) == b"A" * 127  # test cap 127 chars
+
+
+def test_encryption_unicode(tmp_path):
+    "Issue #933"
+    pdf = FPDF()
+    pdf.set_author("Thai")
+    pdf.set_subject("ทดสอบภาษาไทย")
+    pdf.add_page()
+    pdf.set_text_shaping()
+    pdf.add_font("Garuda", fname=HERE.parent / "fonts" / "Garuda.ttf")
+    pdf.set_font("Garuda", size=12)
+    pdf.start_section("ทดสอบภาษาไทย")
+    pdf.cell(
+        txt="สวัสดี ทดสอบภาษาไทย กีกี้ กาก้า ก๋า อ้า อ้ำ ฤาษี ทุ่มทุน อุ้งอุ๋ง น้ำใจ ฯลฯ ญาญ่า ฐาน ฎีกา ฏฒัฯนณ ภัทร์ สิทธิ์"
+    )
+    pdf.set_encryption(owner_password="fpdf2")
+    assert_pdf_equal(pdf, HERE / "encryption_unicode.pdf", tmp_path)
diff --git a/test/fonts/Garuda.ttf b/test/fonts/Garuda.ttf
new file mode 100644
index 000000000..b9fcdfd90
Binary files /dev/null and b/test/fonts/Garuda.ttf differ
diff --git a/test/fonts/charmap_first_999_chars-Garuda.pdf b/test/fonts/charmap_first_999_chars-Garuda.pdf
new file mode 100644
index 000000000..a3c0c4b1c
Binary files /dev/null and b/test/fonts/charmap_first_999_chars-Garuda.pdf differ