A light stripe extraction algorithm based on double-Gaussian fitting is proposed to solve the influence of high-reflective light on a light stripe extraction when the line laser measurement system measures the metal surface. First, the multipeak distribution rule of light stripe gray level is determined by analyzing the gray level of the light stripe section. Second, the reflection model of light is deduced, and the generation principle and energy distribution model of high-reflective light on metal surface are studied. Then, a double-Gaussian fitting model is established based on the distribution model, and a light strip extraction algorithm is designed. The effectiveness of the proposed algorithm is verified using examples. Finally, a comparative experiment is performed to analyze the extraction effect of the double-Gaussian fitting method and the traditional light stripe extraction algorithm and to evaluate and analyze the confidence of the results. The results show that the double-Gaussian fitting method can effectively suppress the influence of high-reflective light in the light stripe image, accurately extract the center of the light stripe, and its confidence evaluation performance is better than that of the traditional algorithm.