It is very challenging to predict a crime scene only by machine without human intervention. This research has tried to make that possible. Convolutional neural network (CNN) has been used to detect 4 objects which are handgun, fire, knife, and accidents. By detecting these objects easily with the help of CCTV cameras, the machine can predict the crime scene. Machines will be able to identify crimes swiftly and intervene based on situations like accidents or violence. This research has adopted a variety of techniques to reach the pinnacle of implementation and success. The model used here has been built with the help of CNN, and there are 4 objects to classify which are mentioned earlier. This research has succeeded in predicting crime scenes through CCTV cameras which may bring prosperity to the country and the nation.