Abstract: In order to enhance the performance of numerical optimization algorithm, a self-adaptive-learnings-based ensemble algorithm (SALBEA) is proposed. In the SALBEA, greedy breeding operator, X-evolution operator, population diversity maintaining operator and evolution strategy learning operator are designed to enhance the evolution structure. Besides, a probability model and a self-adaptive learning mechanism are employed to integrate four effective search strategies. Firstly, in order to evaluate the performance of SALBEA, 26 state-of-the-art test functions are solved by the SALBEA and its competitors, and experimental results indicate that the effectiveness and robustness of the SALBEA outperform its competitors. Then, the SALBEA is employed to solve matrix eigenvalue problem. Experimental results show that the precision of the solutions is very high and SALBEA is a promising algorithm in real-world application.