With the development of modern military strategy, satellite observation has become a key asset for obtaining global security and operational environment dynamics. This article proposes an intelligent scheduling method for ground observation of satellite based on segmented coding, aiming to optimize observation plans, improve the quality and efficiency of data collection. The article first analyzes the current research status of satellite observation duration allocation and points out the shortcomings of current research on ground observation task duration allocation. In response to this issue, this article establishes an observation duration allocation model, which maximizes the observation benefits of satellites in different orbits (high, medium, and low) by adjusting the observation duration decision variables of each satellite. The model introduces an encoding system to represent different observation time allocation schemes and establishes a relationship function between observation duration and benefit value. Furthermore, this article proposes an improved genetic algorithm based on allele replacement to solve the allocation problem of observation time. The experimental results show that the method tends to stabilize after 10000 iterations, and the total allocatable time resource utilization rate of the proposed scheduling scheme reaches 99.9999%, which is much higher than the benefit value of the uniform allocation scheme, proving the feasibility and effectiveness of the proposed method in this paper.