With the development of mobile technology, the popularity and usage of smartphones have increased exponentially, and attackers are constantly seeking to obtain sensitive, confidential information from smartphones. To protect smartphones from malicious applications, researchers are continuously working on effective Android malware detection approaches. Machine learning-based techniques proved to be effective for Android malware detection. This study reviews popular machine learning approaches widely used in static analysis of Android malware detection. The paper’s goal is to find the machine learning approaches that give the best results when used for the classification of benign and malware applications. After a detailed literature review, an experimental setup was designed to test the six most popular Machine Learning (ML) approaches over three datasets of Android malware and benign applications. The observations from the previous studies are also verified by experiment and the results are compared using various performance metrics. Results reveal that the Random Forest (RF) classification algorithm achieved better performance over nine selected metrics when used to classify Android malware and benign applications using static features.