A new load detection method was presented based on its running status of power system. The loads are sorted according to their index of capacitive, inductive, and none-linear. The static character of each load was got according to the time-distribution. The concept of load detection is completed. The data-sample and wave recording equipment is designed and assembled based on the DSP for power loads status detecting. The superior software was designed for data management and analyzing. The recognizing status data was analyzed in both time-domain and frequency-domain; the identifying method for electric power loads status is deduced based on the correlation coefficient; the mathematical model for the status identifying is built; the trueness and validity of the methods is verified based on the testing and analyzing