To enhance global exploration and local exploitation performance of the multi verse optimizer(MVO), the improved multi verse optimizer(IMVO) is proposed by coupling horizontal and vertical individual updated strategies.The horizontal updated strategy is a horizontal migration evolution mechanism on the population level , in which population diversification and global exploration performance can be improved by introducing the weighted learning factor(WLF) to guarantee that offspring individuals inherit position information from multiple universes of parent generation simultaneously.The mathematical expression of wormhole existence probability is amended properly to heighten information exchange among individuals.Moreover, the vertical updated strategy is a vertical self-learning evolution mechanism on individual level, in which local exploitation performance can be modified by simulating the history forgotten memory characteristic of human cognition to ensure that the neighbourhood of memory mean position based on history information of best universes is re-exploited.Experimental results verify the performance influence of difference WLF functions on the IMVO, and the better optimization performance and robustness of the proposed algorithm.