A network intrusion detection model based on the hybrid deep learning algorithm of TCN-CNN is designed to improve the success rate of network intrusion detection. The model integrates the temporal convolutional neural network (TCN) and convolutional neural network (CNN) frameworks to enhance the accuracy of network intrusion detection. First, TCN is used to extract high-frequency behavior features from long-time sequence data, forming a feature matrix by concatenating multiple feature vectors and converting it into an image for CNN convolutional learning. The optimal weight of the hidden layer feature matrix is then obtained, and the powerful image recognition ability of CNN is used to perform category mapping to assist the network intrusion detection system in achieving network anomaly detection. The experimental results show that the model can detect 94.56% of the five different DDoS attacks, which has a higher accuracy and faster convergence rate than other machine learning-based intrusion detection models.