The invention relates to an automatic driving service unloading method based on edge calculation. The method comprises the steps of S1, enabling a vehicle with an automatic driving function to sense the surrounding environment through various sensors and obtains data; S2, querying and detecting a current network state and available resources through an API before the vehicle runs; S3, enabling theedge or the cloud to calculate unloading strategies in different network states through an ILP integer linear programming model, and store the unloading strategies as a lookup table; S4, when the vehicle runs, registering the current network state with the stored lookup table, and when the matching is unsuccessful, performing adaptive dynamic scheduling by using a heuristic algorithm, and takingthe nearest task unloading strategy in the lookup table as the reference of the heuristic algorithm for iteration; and S5, returning a calculation result to the unmanned vehicle by the edge server andthe cloud, and repeating the operation in the step S4 at the beginning of the next period. According to the method, the low-delay performance required by unmanned driving is ensured, the service quality is improved, and the response time of an automatic driving calculation task is effectively shortened.