The traditional odometer correction method often uses off-line correction means, and the correction effect is poor when the ground environment changes. In order to solve the above problems and improve the positioning accuracy of the robot, an online odometer correction method with simple structure and strong robustness was proposed on the platform of three wheeled omnidirectional mobile robot. In this method, the extended Kalman filter algorithm was used to process the sensor data to obtain the robot’〖KG-*2〗s real-time pose information and speed information, and the dynamic model of the three-wheel omnidirectional mobile robot was combined to correct the odometer error in time. A simulation experiment was designed in V-REP. The experimental results show that the odometer accuracy has been greatly improved by using the proposed method, and the problem of offline calibration method affected by the ground environment is overcome. The effectiveness of the algorithm is verified on the existing actual platform.