<p>The imaging satellite task of multi-mode and multi-point targets is widely used in military reconnaissance, environmental monitoring, meteorological forecasting and other fields. To address the multi-task observation conflict and strip sequence decision-making problem under constrained conditions, and to effectively resolve observation conflicts between different task modes, this paper proposes a greedy algorithm based on Heuristic Dynamic Programming (HDPGA). The objective is to maximize task value while minimizing time expenditure, ensuring all tasks are completed with reduced time costs, thereby achieving optimal overall efficiency. Using dynamic programming can decompose tasks into subtasks and store solutions to subproblems to avoid redundant calculations, retain the advantages of greedy algorithms (GA) in efficient computation, and efficiently complete the entire task by solving subtasks. This study first analyzes the demand characteristics of multi-mode and multi-point target tasks, establishes a model to ensure that task constraints are met, then uses a greedy algorithm based on heuristic dynamic programming to solve the model, and finally conducts simulation experiments and result analysis. The experimental results show that when paying for 7 cycles of time cost during single orbit operation, using HDPGA algorithm can obtain all values, while using traditional GA algorithm can only obtain 63.64% of all values. When all tasks are completed and all value benefits are obtained, HDPGA algorithm can reduce time costs by more than 30% compared to GA algorithm.</p>