Abstract: In order to diagnose and evaluate the human spinal lesions through the paravertebral muscles, a paravertebral muscle monitoring system based on surface EMG signals was designed. The system used surface mount electrodes to obtain the surface myoelectric signal (sEMG) of paravertebral muscle. The signal was filtered and amplified by the conditioning circuit. The signal was collected by the microcontroller NRF52832 and was sent to the mobile APP. After the signal was preprocessed by the wavelet threshold denoising algorithm in APP, the time and frequency characteristics of the sEMG signal reflecting the functional state of the muscle were extracted. The calculated characteristic parameters was displayed in real time in the application interface. The experimental results show that the system meets the design requirements in analog signal acquisition, digital processing of signals and calculation of characteristic parameters. The system has certain application value.