Malware detection in online content is a tedious and complex task. A lot of recent methods focus on providing robust solutions for detecting malware to prevent malicious attacks by cybercriminals. These methods can detect a few classes of malware but can not guarantee the detection of all the malware present in the system. In order to ensure full protection, it is essential to have a robust methodology that can detect different types of malware. To this end, we propose an advanced machine learning-based approach exploiting TPOT methodology for different types of malware detection. TPOT is an automated machine learning methodology executing multiple machine learning algorithms in a pipeline. This methodology uses multiple preprocessing techniques to optimize hyperparameters and tune the model for its best performance. Feature selection strategy is adopted to select sensitive features for reducing model computational complexity. Experimental validation of the method on public dataset has proved the robustness of the proposed method in detecting malware.