Low signal-to-noise ratio (SNR) and lack of source information are the main factors leading to early fault inaccurate detection. In order to solve this problem, a dual vector time-time domain (DVTD)method is proposed to complete and highlight the early weak fault characteristics of gears. In this method, the full vector principle is used to realize the fusion of the vertical double channel vibration signals, so as to obtain dual vector signal having more complete source information. On this basis, combined with the time-time domain transform theory, the principal diagonal elements of two-dimensional time series are extracted to construct a complete and fault feature enhanced time-domain vibration signal. Finally, the wind turbine gearbox is taken as the experimental object, and the index at small scale characterizing the fluctuation strength of the signal is extracted as the state feature. The classification results verify the validity of the dual vector time-time domain weak fault feature enhancement method in gear early fault identification.