Wind turbines are often deployed in harsh and dynamic environments, rendering their gearboxes susceptible to unexpected faults that can result in significant downtime. Timely and accurate fault detection is therefore essential to ensure stable operation, minimize maintenance costs, and enhance overall system reliability. In this study, a fault detection framework for wind turbine gearboxes, which is based on a one-dimensional convolutional neural network (1D-CNN), is presented. Three representative fault conditions, including gear tooth breakage, gear tooth corrosion, and combined tooth breakage with corrosion, are constructed for evaluation. Triaxial vibration signals are collected using an accelerometer to capture critical operational features. The proposed 1D-CNN model is trained and tested using time-domain vibration data, enabling automated feature extraction and fault classification. Experimental results confirm the model’s effectiveness in accurately identifying gearbox faults, demonstrating its potential for reliable health monitoring of wind turbine systems.