In order to enable the Interactive Water Sanitation Vehicle (IWSV) to capture the floating garbage and avoid the obstacles in the water smoothly during the garbage collection, a path planning algorithm combined Rapidly-exploring Random Tree (RRT) algorithm based on the sampling with the speed obstacle model is proposed. The position coordinates of dynamic obstacles in the water are obtained based on the parallax positioning method with the binocular camera, the relative azimuth between the sensing element carried by IWSV and obstacles is obtained, the adjustable range of movement angle to avoid the obstacles successfully is calculated based on the speed obstacle method, the random sampling process in the better RRT* algorithm is further optimized, and an improved obstacle avoidance path planning algorithm is obtained. In view of the practical application scenes, the anti-integral saturation Proportional Integral Differential Control (PID Control) method is introduced to make the control effect of course controller more precise and effective. In the real scene test, the obstacle avoidance path planning algorithm is of the robustness, and the simulation analysis is conducted based on Time of Arrival (TOA) positioning method. The simulation results show that the proposed path planning algorithm is of better effect and reliability than the RRT algorithm and the improved RRT* algorithm, and it can avoid the obstacles in a short time and get a better moving path. In the real scene test, the Chan algorithm based on TOA is more accordant with the positioning estimation requirements, and the noise measurement of sensing device of IWSV itself should be within 10 m.