The invention discloses an intelligent monitoring and early warning method for an elevator brake. The method comprises the following steps of 1, a hardware structure of an elevator brake intelligent monitoring and early warning system is built in an elevator machine room; 2, small sample monitoring data are collected, feature vectors of signals are decomposed, reconstructed and extracted through awavelet packet, and a health mode library is built; 3, the running state of the elevator brake is monitored in real time, and early warning is carried out on faults possibly occurring through an SVMclustering analysis algorithm; and 4, according to a fault prediction result, maintenance personnel of an elevator are guided to adjust and optimize the maintenance plan, predictive maintenance is carried out, the faults of the brake are eliminated in a sprouting state, and accidents are avoided. According to the intelligent monitoring and early warning method, key parameters like an elevator brake gap, the abrasion amount of a brake shoe, a brake distance, a brake coil voltage, a traction machine working current, brake temperature, noise are monitored in real time through a non-contact type measuring method, data mining and analysis are carried out on the large amount of monitoring data, so that fault hidden danger of the elevator brake is predicted in time, and safe operation of the elevator is guaranteed.