In order to improve the testing precision and real-time performance of the AC electrical power performance index, the coarse signal processing approach of AC power intelligent sensor was studied based on the theory of correlation analysis method and the least-squares method. The least-squares characteristic parameters method was used to estimate the initial sample points of AC voltage and AC current, and the electrical power parameters were obtained. The formulas of calculating quantity and computation complexity were discussed by comparing the least-squares characteristic parameters method with the correlation analysis method.The results show that when the noise signal amplitude is permitted from 3% to 12%, the proposed method can reduce the processing time by 49.4% and decrease the testing error by 2/3 ,and the requirement for the signal-to-noise ratio of the system can also be reduced as opposed to the traditional method.