Malware attacks continue to pose a significant threat to computer systems and networks worldwide. Traditional signature-based malware detection methods have proven to be insufficient in detecting the increasing number of sophisticated malware variants. This has led to the exploration of new approaches, including machine learning-based techniques. In this paper, we propose a novel approach to malware detection using residually connect convolutional networks. We demonstrate the effectiveness of our approach by training CNN on a large dataset of malware samples and benign files and evaluating its performance on a separate test set. Extensive experiments on a public dataset of malware images demonstrated that our model could achieve high accuracy in detecting both known and unknown malware samples. The findings suggest that our residual convolution has great potential for improving malware detection and enhancing the security of computer systems and networks.