To address the problem of inaccurate identification of the receiving end caused by background light interference on the transmitting light source in underwater wireless optical communication, an underwater laser spot circle fitting center positioning algorithm is proposed based on a random sampling consensus algorithm. First, the spot image is preprocessed to obtain edge pixels. Then, a random sampling consensus algorithm is used to eliminate the interference of noisy data points, and finally, a circle fitting algorithm is used to position the spot center with high precision. Experimental verification was conducted by shooting laser spot images in water in the presence of background light interference. The results show that compared with the traditional centroid method and circle fitting algorithm, the improved algorithm significantly reduces the impact of background light on the positioning of the light source center. In the x direction, the positioning error was approximately 2 pixel, and the minimum error was only 0.49 pixel; in the y direction, the error was approximately 1 pixel, and the minimum error was only 0.04 pixel. The average coordinate deviation was 1.805 pixel, and the root mean square error for the calculated spot center coordinates was 2.2063 pixel, demonstrating the low deviation, high accuracy, and good anti-interference ability of the improved algorithm.