This paper proposes an improved sparrow search algorithm with multi strategy fusion (ISSA) to address the issues of dependence on initial population distribution, susceptibility to local optima, and reduced population diversity in the later stages of iteration in sparrow search algorithm (SSA). Firstly, the population is initialized using Sobol sequences to ensure the diversity of the initial population; Secondly, random reverse learning strategy and spiral foraging strategy are introduced to improve the discoverer position update formula and the joiner position update formula, respectively, to enhance the algorithm's global search ability and ability to jump out of local optimal solutions; Finally, introducing Cauchy variation to perturb sparrows that may fall into local optima. Nine standard test functions were selected for performance testing in the experiment, and the results showed that the improved algorithm had a significant improvement in performance. Applying ISSA to AGV (Automated Guided Vehicle) path planning can achieve optimal values of 13.1356, 28.8345, and 44.3649 in three map environments, respectively. The optimization ability and stability of the algorithm are significantly improved compared to the original algorithm.