The invention discloses an automatic identification algorithm for microseismic signals and p wave first arrival at an engineering scale. The automatic identification algorithm comprises the steps of: reading data of sampling points in a set time window; calculating feature functions as well as long and short time average values of the sampling points; carrying out automatic pickup judgment on the microseismic signals and the p wave first arrival; clearing buffer memory of parameter values to zero; and reading data in a time window at a next period. According to the automatic identification algorithm, the identification algorithm in the seismic field is improved aiming at microseismic signal characteristics at the engineering scale, the automatic identification algorithm can automatically pick up microseismic signals especially weak signals with low signal-to-noise ratio at the engineering scale accurately, increases the accuracy of automatically picking up p wave first arrival, and further enhances early warning timeliness and accuracy of geological disasters such as rockburst, mine earthquake and collapse.