Malicious applications provide a growing and significant threat to Android users, developers, and application platforms. Experts have endeavored to create novel detection methodologies due to the ongoing advancement in the sophistication of malware and the escalating intensity of its damaging assaults. Within the framework of these initiatives, the detection of malware encounters a significant impediment due to the lack of clean and balanced datasets. This research seeks to develop a model proficient in identifying and categorizing various forms of hybrid malware on the Android platform, utilizing the Decision Tree (DT) model as the primary analytical instrument. The purpose of study is employing Machine Learning (ML) techniques for the categorization of network traffic related to malware programs. It presents ML approaches utilizing DT, K-Nearest Neighbors (K-NN), Naive Bayes (NB), and Logistic Regression (LR) for predicting network virus traffic. It conducts experiments on the CICAndMal2017 dataset. The technique comprises many essential stages, such as data processing, which includes the raw data that is refined through cleansing and transformation from text format to numerical format. To address the problem of class imbalance among application categories, the oversampling approach was employed to ensure equal representation of all malware types in the dataset, followed by feature engineering. The features were assessed via the Random Forest (RF) model to determine the permissions and behaviors that most significantly impact application classification. This facilitated comprehension of the principal variables behind harmful acts. The findings indicated that the DT model was proficient in prediction and showed superior performance relative to other models. The last classification reports confirmed that the model achieved a good balance between precision and recall in the classification of both secure applications and various malware types. The model's performance was measured by its accuracy, F1-score, recall, and precision; it achieved a score of 99.99% in all measures utilized