An improved ant colony algorithm (ACA) was adopted for the dynamic path planning problem of mobile robots (MR). This algorithm combines the advantages of ACA and artificial potential field method (APFM), and introduces the potential field function from APFM into ACA, utilizing the relationship between the resistance of ant movement and movement speed. Finally, simulation experiments were conducted on the algorithm, and the results showed that the planning efficiency of the improved ACA was as high as 89.1%. Compared with traditional ACA, it can find the optimal path faster, reduce search time, and also improve the path planning ability of the robot in dynamic environments.