Aiming at the characteristics of various types of equipment in coal preparation plant and the dispersion of monitoring points, a multi-source information wireless transmission and classification algorithm for equipment based on compressed sensing was proposed. By constructing a multi-hop information transmission model, the information transmission problem was transformed into the compressed sensing problem of multi-path measurement signals, thereby the measurement matrix acquisition was transformed into the routing problem of the multi-hop information transmission model. Aiming at the large coherence of the obtained measurement matrix and affecting the signal reconstruction effect, the idea of random routing was introduced into the routing construction, and a random dynamic self-organizing routing algorithm was proposed. In order to solve the problem that the time domain features of the reconstructed signal were difficult to accurately classify the fault type, a new time domain feature, the total variation (TV) of the vibration signal, was introduced for the reconstructed signal, and the compensation distance estimation algorithm was adopted to verify the superiority of the introduction of indicators. The analysis of the measured data of the coal preparation plant shows that the proposed multi-source information transmission and classification algorithm can effectively improve the fault recognition accuracy under the condition of improving the real-time transmission efficiency of the monitoring data.