Android malware recognition is the process of detecting and preventing malicious software that is created with Android OS in mental state. (OS), which are widely used in tablets and smartphones. One of the most common types of cyberattacks is malware, which is becoming more common every day across the network. These vulnerabilities make it easy for a hacker to obtain the private information on a mobile device. To overcome this issue a novel Attention Based Cnn-bi-lstm NETwork based malware detection (ABC-NET) has been suggested to solve this issue by precisely identifying and thwarting malware attacks and enhancing device security. The first step is to collect the data that was gathered from the Android. Information based on actions and information based on endorsements feature extraction are the two types of data. After being converted into sequence data, the extracted features are fed into the classification step. The CNN-BILSTM attention-based technique is used to differentiate between benign and malicious data during the classification phase. The suggested approach outperforms current methods like DIEL (89%), OEL (91%), and GAN (94%) in identifying and reducing malware threats on Android devices, with an overall accuracy of 98%. This illustrates how much more effective the suggested model is than more conventional deep learning techniques