With the introduction of augmented reality (AR) technology into mobile devices, it becomes a trend to develop mobile AR applications in various fields. However, the limited mobile hardware resources, like the CPU frequency, memory capacity, etc., makes it difficult to guarantee the delay of resource-intensive AR applications. In response to this challenge, we propose a mobile AR offloading method under the edge-assisted environment. Firstly, we divide an AR task into multiple consecutive subtasks, and then collect the features of hardware, software, configuration, and runtime environments of edge servers to be offloaded. With the features, we construct an AR subtask Execution delay Prediction Bayesian Network (EPBN) to predict the execution delay of different subtasks on each edge platform. Based on the prediction, we model the AR task offloading as the NP-hard Traveling Salesman Problem (TSP), and then propose a PSO-GA based solution by adopting the heuristic algorithm of Particle Swarm Optimization (PSO) to encode the offloading strategy and using Genetic Algorithm (GA) for particle update. The extensive experiments prove that our proposed method can optimize the AR task offloading strategy with the lowest delay and outperform the other baselines.