An optimization strategy for multi-layer edge computing networks based on computing power and channel quality was proposed to address the problem of wireless resource management caused by communication and computing power heterogeneity in multi-layer edge computing networks. A performance analysis of a typical multi-layer edge computing network was presented, including communication performance and edge computing latency performance. A mixed-integer nonlinear programming (MINLP) problem was formulated based on the computing power and transmission power of base stations, the computing power and task characteristics of devices, as well as the user association, offloading decision, and resource allocation, aiming to minimize the latency of local computing and the upload, processing, and backhaul latency of edge computing. Based on an iterative optimization approach, the optimal user association was achieved by considering both channel quality and computing power, a task offloading strategy was developed to balance local and edge computing, and wireless resource allocation was optimized to minimize edge computing latency. Simulation results highlighted the importance of jointly optimizing spectrum and computing resource allocation, as well as optimizing task offloading in edge computing networks, and showed that user association optimized based on both channel quality and computing power achieved lower latency compared to traditional methods that relied on only channel quality or computing power.