The drastic modernization in computer era has brought a revolution impact in the medical science field. In medical science the designing and screening of drug plays a vital role. The computer simulation is been used for preparing computational medicines or drugs which are known as silico drug. The most popular computer simulations are Artificial Intelligence and its essential components (Machine Learning and Deep Learning). Artificial Intelligence and its essential components are more accurate, cost effective and shorten the drug designing and screening process. So, that infection can stop spreading quickly. The earlier Artificial Intelligence methods faced with many challenges such as unable to handle the imbalanced data, efficient deep learning methods were not considered and complexity (time and space) of earlier algorithms were high. To overcome the challenges of earlier methods a new algorithm Medicinal Drug screening using Residual Convolutional Neural Network (MDSRCNN) is been proposed in this paper which tries to overcome earlier algorithm problems. In this review the structure of medicinal drug, steps of medicinal drug development, earlier artificial intelligence methods and its challenges, our proposed algorithm MDSRCNN and finally, the evaluation measures like precision, recall, F1 and accuracy are been considered.It is been concluded that the proposed MDSRCNN algorithm shows a good accuracy of 99 percentage and is comparatively accurate than other previous methods.