Multiple dynamic targets are often affected by factors such as changes in lighting, mutual occlusion between targets, and dynamic backgrounds, resulting in poor tracking performance. To solve this problem and achieve accurate and timely tracking of multiple dynamic targets, real-time remote sensing technology is utilized to optimize the design of real-time tracking algorithms for multiple dynamic targets. The article first sets the standard features for tracking targets based on their structure and color characteristics. Utilize real-time remote sensing technology to dynamically obtain target remote sensing images, and achieve remote sensing image preprocessing through steps such as filtering, defogging, and occlusion processing. Then, the background subtraction method is used to segment multiple dynamic targets in remote sensing images, extract features of multiple dynamic targets from both structural and color aspects, and finally achieve recognition and localization of multiple dynamic targets through feature matching. By estimating the motion of multiple dynamic targets, the tracking results of all dynamic targets at any given time can be obtained. The actual test results show that compared with traditional algorithms, the optimized design algorithm significantly reduces tracking errors and increases update frequency in global lighting and occlusion scenarios. It has higher tracking accuracy and real-time performance, and has good application effects.