Abstract: The gearbox is a key component of the drilling rig transmission system, and its operating condition directly affects the efficiency and safety of equipment operation. To address the problems of non-stationary vibration signals, difficulty in feature extraction, and insufficient recognition accuracy under small-sample conditions, an intelligent fault diagnosis method based on the wavelet synchrosqueezing transform（WSST） and CNN-BKA-LSSVM is proposed. First, WSST is used to perform time-frequency analysis on the original vibration signals, thereby enhancing the separability of fault features. Then, a convolutional neural network（CNN） is employed to automatically extract deep features from the time-frequency images. Finally, the black-winged kite algorithm（BKA） is adopted to optimize the key parameters of the least squares support vector machine（LSSVM） for fault classification. Experiments are carried out on the gearbox of a ZDY1800G fully hydraulic tunnel drilling rig, with fault types including healthy condition, broken tooth, missing tooth, tooth root crack, and surface wear. The results show that the proposed model achieves a classification accuracy of 99.5% on the test set and maintains good stability and generalization ability under small-sample conditions, outperforming comparison models such as CNN-LSTM-Attention and VMD-CNN-BiLSTM. The findings can provide a useful reference for fault diagnosis of drilling rig gearboxes and other coal mine machinery.