In the wheat harvest period, when the weather changes, the wheat harvesting time window will change, if the original scheduling scheme is not adopted in a timely and efficient dynamic scheduling method, it will cause a large number of farmers to delay the harvest of farmland, which will bring huge losses to farmers. To avoid or reduce the loss of farmers caused by weather change, considering the emergency scheduling problem of farmland harvester time window change caused by weather change, an multi-objective emergency scheduling model considering the transfer time, waiting time and delay time of the harvester was established, it considered a shortened harvest time window caused by weather change, and an emergency degree function of farmland harvest was introduced to determine the emergency farmland. The farmland with the intersection of working time and rainy time would be prioritized according to the emergency degree. The genetic algorithm was improved according to the characteristics of this model by using two-level multi segment coding method and single point crossing method, so that the algorithm can effectively avoid local optimization and achieve global convergence. Based on the improved genetic algorithm, an emergency harvester scheduling algorithm was designed to achieve the priority of emergency farmland by changing the original harvester route dynamically, so as to improve the service capacity of agricultural machinery cooperatives and reduce farmer losses. Finally, an example was introduced to prove the feasibility of the model and algorithm.