Abstract: The conventional method for positioning coal mining machines predominantly depends on the inertial navigation system. In the challenging operating conditions of a fully mechanized mining face, the positioning system is susceptible to cumulative errors, leading to a progressive decline in accuracy. Additionally, the coal mining machine occasionally encounters irregular driving events, such as slippage, causing anomalies in odometer readings. To address this issue, this paper presents an integrated positioning approach utilizing robust cubature Kalman filtering. By loosely coupling the SINS and OD systems, the speed and displacement data from the odometer are merged with the inertial navigation system’s solution results and incorporated into the filtering algorithm as error observations. Simultaneously, a robust processing method based on the chi-square test of innovation is introduced into the filtering algorithm. Through the construction of a variance inflation factor and adjustment of the observation covariance matrix, it influences the cubature Kalman filtering procedure to achieve robust filtering. The filtered output is fed back into the system as error correction to enable feedback adjustments, and the effectiveness of the combined state is evaluated once the entire solution process is finalized. Simulation and experimental results show that the algorithm can effectively reduce the influence of gross errors, and the positioning accuracy can reach 0.1m after using the integrated navigation algorithm.