There is often a coaxiality error between the rotary part mounted on the rotary table of the machine tool and the turntable, which affects the machining accuracy of the part. In the detection of coaxiality, the detection of eccentricity is a key issue. This paper reviews the current eccentricity calculation methods and proposes a fast calculation method of eccentricity based on least squares fitting(hereinafter referred to as LSF). Different from LSF, this method does not need the absolute coordinate values of the contour points on the rotary part, but only the relative displacement of the contour point relative to the distance measuring instrument. In addition, the method uses four points on the contour of the part as a set of data and N sets of data in total to calculate. The simulation verifies the accuracy and feasibility of the proposed method. Then, analyzes that the method is suitable for the case where the ratio between radius and eccentricity is greater than 200. The comparison proves that the method is 40-50% faster than the calculation of LSF. In order to further illustrate the feasibility of the algorithm, this paper carried out experimental verification. The results in this paper show that the method derived in this paper can quickly and accurately calculate the eccentricity between the rotary parts and the turntable.