The paper presents a new methodology for monitoring health conditions of mining conveyor belts using Radio Frequency Identification (RFID) based sensors. The existing monitoring technique is based on simple visual inspections which is quite labor intensive and do not provide accurate health condition of the conveyor belt. The new methodology based on UHF chipped and chipless RFID sensors provides highly sophisticated real-time monitoring schemes for different conveyor belt health parameters such as cracks. When combined with machine learning algorithm-based approach, the proposed technique can detect and predict wear and tear over the entire belt. Simulation results show that the proposed methodology offers high accuracy in detecting cracks with width of 0.5 mm and demonstrates the efficiency of RFID sensors to track the crack orientation in the belt.