Compared with calibration, the self-calibration does not require a standard artifact with higher precision level than that of the object to be calibrated, which can achieve the calibration of system errors of precision measuring instruments at a low cost of conditions. The standard artifact in 2D self-calibration usually employs traditional three-positions combination of initial position, relative rotation and translation. The positions are independent of each other, and all position transformations are based on the initial position. The utilization of the hybrid position combined with rotation and translation can void the redundant operation of returning the standard artifact to initial position between two position transformations to simplify the process of 2D self-calibration. The 2D self-calibration based on hybrid position is proved by simulation experiments to be effective in separating the errors of the calibrated objects in the simulated noisy environment, and its uncertainty is equal to that of the traditional combination of non-hybrid positions. The introduction of hybrid position has no influence on the noise suppression capability of 2D self-calibration. The 2D self-calibration experiments are conducted on a vision measuring machine, and the stage system errors separated by the 2D self-calibration using a three-positions combination including a hybrid position are close to that of traditional three-positions combination. The results of the stage system errors corresponding the two positions combinations differ by 0. 167 μm. In addition, the effect of two-dimensional self-calibration based on hybrid position is further demonstrated by experimental results using a four-positions combination including a hybrid position