The Wind Turbines (WTs) experience gradual wear of gears and bearings under variable weather conditions.This paper presents a novel fault diagnosis method for the drive train of WTs based on Adversarial Discriminative Domain Adaptation (ADDA) and Acoustic Emission (AE). AE, serving as a non-destructive testing technology, is used to detect rub-impact faults in rotating machinery. ADDA, as a form of transfer learning (TL), takes knowledge from a source domain and applies it to a target domain, thereby improving the model's generalization capability. Experiments show that the proposed AE-ADDA method achieves excellent fault diagnosis results.It offers a novel approach to tackle the challenges of fault detection in wind turbine drive trains across varying operational conditions.