In current space-based reconnaissance task pretreatment process, resource allocation generally lies on task requirements, which may lead to load imbalance. In light of this, a matching method that considers both task requirements and resource load was presented. For the sake of load balancing, the key task elements and their description method were analyzed. Based on the task flow model, formula and algorithm for calculating the probability of the executed task were available, then the execution probability were used to measure the load status of satellite resources. Finally, the research put forward a load balancing model and a corresponding heuristic algorithm for matching reconnaissance tasks and satellite resources, which takes into consideration the workload, priority, quality of tasks and execution probability. Simulation results show that the proposed load balancing method can effectively improve the utilization of satellite resources and reconnaissance benefit.