Based on the noise analysis of the photoelectric detection circuit, error compensation aiming at photoelectric detection circuit has been performed by making use of interpolation method, look-up table method, adaptive neuro-fuzzy inference method, and surface fitting method, so as to improve the accuracy of photoelectric detection as well as reducing the system error. In the meanwhile, the error compensation effects of different methods are compared and analyzed. Finally, with the help of the MATLAB software, these effects are simulated using the measured data of photoelectric diode experimental platform as the sample data. The result of the simulation indicates that the adaptive neuro-fuzzy inference method can improve the detection accuracy better and its error compensation effect is also better than that of interpolation method, look-up table method, and surface fitting method, when they are under the same conditions.