A light stripe center extraction algorithm is proposed based on Gauss-Lorenz peak fitting, which is suitable for different types of light stripe images. The cross-sectional energy model of a light stripe is built and the cross-sectional energy compositions of light stripes with different image qualities are analyzed, thus the Gauss-Lorenz decomposition fitting model is constructed. Based on the constructed model, the successive gray normalization, extraction of regions of interest, Gauss-Lorenz decomposition peak fitting and removal of Lorentzian components are acted on the light stripe images. The accurate light stripe center can be extracted with the gray centroid method. The contrast experimental results show that the proposed method has the characteristics of high extraction accuracy, strong applicability, but long time-consumption for different types of light stripe images. However, if it is applied to strong diffuse reflection and specular reflection light stipes, its time consumption is only half of that of Steger algorithm with the best effect at present.