Unmanned Aerial Vehicles (UAVs) are an emerging novel type of equipment that accomplish predetermined goals through task preplanning by the base station (BS). However, with the increasing application of UAVs, they face challenges such as changing task environments, high communication pressure on BS, and high time complexity of global planning algorithms during task execution. To address these issues, collaborative task replanning for multi-UAV is required. Therefore, this paper proposes an Incremental Task Replanning Algorithm (ITRA) for multi-UAV based on centralized-distributed negotiation to solve the problem of dynamic task planning. By establishing periodic communication of UAVs in a centralized-distributed mode, the real-time sharing of UAV information and self-organizing management are achieved, enabling multi-UAV to quickly organize autonomously to adapt to environmental changes. Based on this, we designed a centralized-distributed negotiation incremental replanning model. The model seeks feasible solutions with a minimal scheduling scale by gradually increasing the number of UAVs involved in replanning, avoiding extensive scheduling. Experimental comparisons and validations demonstrate that this method effectively handles unexpected situations in dynamic environments and achieves higher efficiency compared to global replanning under the same conditions.