Abstract: To improve the optimization ability of the golden eagle optimization (GEO) algorithm in solving indoor visible light positioning problems, a new cross-entropy golden (CE) eagle optimization (CEGEO) algorithm was proposed based on the importance sampling technique and the Kullback–Leibler distance CE method.Integrating the cross entropy method into the golden eagle optimization (GEO) algorithm, the sampling probability distribution parameters were updated by a new population obtained through collaborative evolution to accelerate the iterative process of the algorithm, and reduce the number of sampling samples and computational costs, and the global optimization capability of the algorithm was improved. Meanwhile, better individuals were obtained through common updates to significantly increase the diversity of the population, and improve the convergence speed of the algorithm. Simulation experiments on standard test functions and indoor visible light positioning optimization problems were conducted using CEGEO algorithm and four other algorithms to compare and analyze the performance of CEGEO algorithm.The results show that compared with the other four algorithms, the proposed CEGEO algorithm has higher solving accuracy and faster convergence speed, can effectively balance global and local search capabilities, and has stronger optimization ability. In indoor visible light positioning, the positioning accuracy is less than 2.74E–12 cm, the estimated position error is the smallest, and the average error is close to 2.07E–15 cm, which is an effective algorithm for solving the optimization problem of indoor visible light positioning.