Pembelajaran Kanji tingkat N5 dan N4 sering dianggap sulit karena pembelajar harus menghafal bentuk karakter, arti, dan pelafalannya secara bersamaan. Selain itu, media pembelajaran yang tersedia umumnya masih bersifat pasif dan belum mampu menyesuaikan tingkat kesulitan sesuai kemampuan pengguna. Penelitian ini bertujuan mengembangkan game edukasi Kanji berbasis Android dengan mengintegrasikan Speech Recognition dan AI Adaptive menggunakan metode Fuzzy Logic Sugeno. Metode
pengembangan yang digunakan adalah Game Development Life Cycle (GDLC) dengan
implementasi menggunakan Unity dan bahasa pemrograman C#. Speech Recognition
digunakan untuk mengenali pelafalan Kanji yang diucapkan pemain, sedangkan Fuzzy
Logic Sugeno digunakan untuk mengatur tingkat kesulitan permainan secara adaptif
berdasarkan performa pemain. Pengujian sistem dilakukan menggunakan Black Box
Testing, White Box Testing, dan User Acceptance Testing (UAT). Hasil penelitian
menunjukkan bahwa sistem berhasil mengimplementasikan fitur pengenalan suara dan
adaptive difficulty sesuai rancangan. Berdasarkan hasil UAT yang melibatkan 30
responden, aplikasi memperoleh nilai penerimaan pengguna sebesar 80,8% yang termasuk
dalam kategori Baik. Hasil tersebut menunjukkan bahwa game yang dikembangkan
mampu mendukung proses pembelajaran Kanji secara lebih interaktif dan adaptif serta
layak digunakan sebagai media pembelajaran berbasis Android.
Kata Kunci: AI Adaptive, Android, Fuzzy Logic Sugeno, Game Edukasi, Kanji, Speech Recognition.

Learning Kanji at the N5 and N4 levels is often considered difficult because learners must
memorize character forms, meanings, and pronunciations simultaneously. In addition,
existing learning media are generally passive and unable to adapt difficulty levels according
to user abilities. This study aims to develop an Android-based Kanji educational game by
integrating Speech Recognition and Adaptive AI using the Fuzzy Logic Sugeno method.
The system was developed using the Game Development Life Cycle (GDLC) method and
implemented with Unity and C#. Speech Recognition was used to recognize Kanji
pronunciations spoken by players, while Fuzzy Logic Sugeno was applied to adjust game
difficulty adaptively based on player performance. System evaluation was conducted
through Black Box Testing, White Box Testing, and User Acceptance Testing (UAT). The
results show that the system successfully implemented voice recognition and adaptive
difficulty features as designed. Based on UAT involving 30 respondents, the application
achieved a user acceptance score of 80.8%, which falls into the Good category. These
findings indicate that the developed game provides a more interactive and adaptive Kanji
learning experience and is feasible to be used as an Android-based learning medium.
Keywords: Adaptive AI, Android, Educational Game, Fuzzy Logic Sugeno, Kanji, Speech Recognition