Abstract: To enhance the mental map preservation in multilayer network visualization and reduce the cognitive load of readers in processing cross-layer analysis tasks, a force-directed based layout algorithm is proposed for multilayer network visualization. Firstly, the unidirectional attraction coefficient between node copies is calculated; Then, according to the unidirectional attraction coefficient and the current position of node copies, the ideal position and distance difference is calculated; Finally, the loss function is constructed by combining the loss term of distance difference, the energy function of KK algorithm and the loss term of spacing between nodes. The experimental results on ten public datasets including seven social networks and three genetic networks show that compared with other similar algorithms, the proposed algorithm has better performances in the sum of absolute differences of nodes, which indicates the proposed algorithm can significantly enhance the mental map preservation in multilayer networks.