Effective monitoring of mining machinery is of great significance. Sensor nodes, which form the basis of the mine’s digital twin system, often face issues of poor sustainability. Therefore, this study introduces a self-powered triboelectric nano-sensor (STNS) enabled digital twin for self-sustained machine monitoring in smart mine. The STNS is designed with three mutually perpendicular sensor units to ensure responsiveness to vibrational energy sources from different directions. Compared to conventional spring-assisted triboelectric nanogenerator (TENG) structures, it exhibits higher frequency adaptability and bandwidth. For a 2 mm amplitude, the STNS responds to frequencies above 10 Hz, with a frequency linearity error rate of less than 0.05%. Utilizing deep learning, the STNS detects various vibrational parameters with an accuracy of ±1 Hz for frequency and ±1 mm for amplitude. A real-time monitoring system based on a deep learning model was constructed and successfully demonstrated for real-time monitoring of amplitude, frequency, and tilt angle. With STNS installed on vibration motor, real-time recognition of the five operating states of the vibration motor and real-time digital twin monitoring were realized. By large-scale distributed deployment of STNS devices, a self-sustained smart mine digital twin ecosystem can be constructed at a lower cost.