Knee adduction moment is a key parameter that links with the severity of knee osteoarthritis. However, assessment of the knee adduction moment is commonly implemented through the stationary measurement systems in a gait laboratory. The purpose of this study is to develop a wearable sensor system that can be used to estimate the knee adduction moment. A wearable sensor sock, composed of six pressure sensors, were developed using the pressure-sensitive electric conductive rubber. Based on the sensor sock measurements and the reference knee adduction moment obtained from the motion capture system (Vicon), we trained a neural network model to estimate the knee adduction moment. In our validation experiments on healthy subjects, the knee adduction moment can be accurately estimated with the trained neural network model.