This research introduces a reliable route prediction model that leverages deep learning techniques to enhance network routing security. By employing deep convolutional neural networks (CNNs), the model efficiently predicts network routing reliability based on the characteristics of routing events. The collection of routing data is initially pursued on a large scale in the real network environment, and further, the collected data is preprocessed and enhanced. After that, a multilayer CNN model is created for training and testing different routing risk level events. This series of experiments aims to assess the effectiveness of the model presented in this paper under various network and load conditions. The objective is to explore the feasibility of utilizing these methods in real-world network security management scenarios. These experimental results show that the proposed model can mark high risk or low risk for the routing events and prove the feasibility of deep learning technology in improving the routing prediction in computer networks.