The weed mapping UAV swarm is studied and a cascade failure control algorithm based on state switching is designed. First, we analyze the agricultural mapping UAV swarm's characteristics of the working environment, then the single-machine node state of the weed mapping UAV cluster is divided into three types: Primary state, intermediate state, and advanced state. Secondly, a set of failure control algorithms are designed based on the principle of single-node state switching and minimum load. Finally, the validity and optimal operating conditions of the algorithm are verified by numerical studies. The results show that: Different failure processes have different degrees of influence on the structural stability and functional stability of the cluster, and the network is most affected when the intermediate node fails initially. The failure control algorithm has the most significant effect when the high-state node fails; The mapping radius of a certain node and the number of low-state nodes contained in the cluster are the most important factors affecting the cluster mapping. The two factors are positively correlated with the mapping area of the UAV swarm and the former always has a more significant effect on the improvement of the monitoring area.