As the space environment becomes increasingly complex, the demand for space situational awareness using space-based optical sensor networks is becoming more urgent. Existing centralized and distributed sensor scheduling methods cannot meet the dynamic and time-sensitive demands for space target observation. The multi-master sensor scheduling approach based on regional grouping is a feasible solution meeting this need, but conflicts may arise when multiple masters make task allocation decisions simultaneously. To resolve conflicts in the task allocation process, we developed a post-allocation conflict resolution method based on the contract net with confirmation protocol, and a pre-allocation conflict resolution method based on the contract net with preselection protocol. Additionally, we have developed two heuristic preselection rules: the revenue-priority rule and the scarcity-priority rule. Experimental results show that the combination of the contract net with preselection protocol with the revenue-priority rule has the lowest communication cost and the highest allocation success rate, making it the most recommended conflict resolution method.