卫星光机载荷的光学效能与热设计密切相关，热控系统的模型修正是热设计必不可少的环节。近年国内外涌现出许多基于深度学习和寻优算法的提高热模型修正效率和精确度的方法可供参考，目前还没有进行系统的归纳。本文针对新出现的修正方法进行了总结，重点分析了在卫星光机载荷热控模型修正这一特殊问题中提高修正效率的几种手段——合适的寻优算法、构建代理模型和开发自动修正工具等。具体分析了这三种手段各自的研究进展及其适用条件和局限性，并提出了对于修正工具开发的思考。最后，对卫星光机载荷热控模型修正领域进行了展望，为后续提高热模型精确度和修正效率提供了方向。

The optical efficiency of satellite optomechanical loads is closely related to thermal design, and model correction of their thermal control system is an essential part of thermal design. In recent years, many reference based methods for improving the efficiency and accuracy of thermal model correction based on deep learning and optimization algorithms have emerged both domestically and internationally. However, there is currently no systematic induction. This paper summarizes the new correction methods, and focuses on the analysis of several means to improve the correction effi-ciency in the special problem of satellite optical mechanical load thermal control model correction - appropriate optimi-zation algorithm, building Surrogate model and developing automatic correction tools. A specific analysis was con-ducted on the research progress, applicable conditions, and limitations of these three methods, and suggestions were put forward for the development of correction tools. Finally, prospects were made for the field of satellite optical and mechanical load thermal control model correction, providing direction for improving the accuracy and correction effi-ciency of thermal models in the future.