In the complex network environment, the existing low-latency transmission technology of HD video data is easy to encounter the problems of jitter and packet loss during HD video transmission, and these problems will directly affect the fluency and clarity of the video, thus reducing the security and quality of transmission. Therefore, a method of low latency secure transmission of HD video data based on 5G communication network is proposed. First, the best matching block and motion vector of the current frame macro block in the previous frame are found through the motion estimation, and then the generated prediction block is compensated and the residual difference is calculated. The prediction block and the residual block are reorganized to recover the video frame to obtain the compressed video data. Then, combined with the QoS requirements, we define the priority of video key slices and ordinary slices, and dynamically allocate resources to different levels of slices according to the network transmission resources. Finally, combined with edge nodes and cloud servers, the video data transmission model is built with the support of 5G communication network. The test results show that the bandwidth occupancy rate is low and the safe transmission quality is good, which has a good application value.