Abstract: This paper proposes a multi-strategy Improved Marine Predator Algorithm (IMPA) to address the Marine Predator Algorithm’s (MPA) problems of slow convergence, low convergence accuracy, and tendency to fall into local optimum. The proposed algorithm introduces a logistic chaotic mapping to initialize the population and increase the diversity of the predator population; an adaptive moving step dynamic adjustment strategy based on the current iteration number t enhances the ability of the algorithm to escape from local optimum; a mid-pipeline algorithm (MA) is added in the late iteration of IMPA, and a free particle position update method based on the mid-pipeline strategy accelerates the position update of the predator and enhances the algorithm’s accuracy and optimization search speed, avoiding falling into local optimum. Finally, the search process is further balanced and the global and local adaptability is enhanced by changing the IMPA stage to transform the search process. Six benchmark test functions are selected to test the performance of the algorithm. Our test results show IMPA converges faster and achieves a higher convergence accuracy. Finally, the improved algorithm is applied in mobile robot path planning, and the simulation results show the algorithm plans a shorter path and achieves a higher search efficiency.