The Android operating system has revolutionized the mobile industry and society with its versatility and customization options. However, the increasing popularity of the Android system has caused a rise in the number of malware attacks, making it challenging to combat these threats. This study proposes a malware detection framework for Android systems that combines permissions and API calls as features. An autoencoder is used to reduce the feature set’s dimensionality, and an ensemble model is employed for classification. The results show that the ensemble models outperform individual classifiers, achieving an accuracy of 98.99%. The study concludes that combining multiple classifiers and feature sets can lead to more accurate and reliable malware detection, providing a foundation for future research in the field of Android malware detection.