In view of the problems of traditional fault diagnosis algorithms based on manual experience and poor data mining,the T1FL fuzzy evaluation method was used to divide the comprehensive evaluation of gearbox system into four health states.In order to scientifically empower the health weight occupied by all levels of state indicators,the digital weight calculation model of least squares coupled enhanced analytic hierarchy process method (AHP) and the improved entropy weight method NBN-EWM was constructed.The results show that the proposed method comprehensively evaluates the gearbox′s health state through the maximum membership principle with the input of multi-source data.A broken gear condition within a 0.920 1 confidence space at the high-speed shaft of the gearbox component is diagnosed,and the evaluation accuracy for the four health statuses is 99%,99.2%,98.2% and 99.6%,respectively.The proposed method can be applied to the early fault diagnosis of gearbox,quantify the health status of gearbox components,accurately locate the damage parts of gearbox,which can provide security guarantee for unit operation and maintenance strategy,and provide new solutions for energy digitalization,intelligent health and operation and maintenance of wind field.