This paper describes the development of a fault detection and diagnosis method to automatically identify different fault conditions of a hydraulic blade pitch system in a spar-type floating wind turbine. For fault detection, a Kalman filter is employed to estimate the blade pitch angle and valve spool position of the blade pitch system. The fault diagnosis scheme is based on an artificial neural network method with supervised learning that is capable of diagnosing a predetermined fault type. The neural network algorithm produces a predictive model with training, validation and test procedures after the final performance evaluation. The validation and test procedures of the artificial neural network model are conducted with the training model to prove the model performance. The proposed method is demonstrated in case studies of a spar floating wind turbine with stochastic wind and wave conditions and with consideration of six different types of faults, such as biases and fixed outputs in pitch sensors and excessive friction, slit-lock, wrong voltage, and circuit shortage in actuators. The fault diagnosis results from the final performance evaluation show that the proposed methods work effectively with good performance. (C) 2021 Elsevier Ltd. All rights reserved.