Power curve (PC) monitoring can be applied to evaluate the wind turbine generator (WTG) power output and detect deviations between the expected and the measured value, often a precursor of unexpected faults. In this research, the instantaneous SCADA data is used to show the fault forecast ability of Artificial Intelligence (AI) based PC monitoring of a pitch regulated WTG. The measured PCs illustrate that the instantaneous data is better than averaged data, widely used in the literature, to present the dynamics of WTG operation. The influence of ambient temperature, generator speed and pitch angle on WTG power output is analyzed using measured data. The analysis illustrates that the generator speed and pitch angle have a significant effect on WTG power generation. The performance of the proposed model option is compared against previously published option using the same data sets collected from a 2 MW Pitch Regulated WTG. The comparison is based on the mean absolute error (MAE), the root mean squared error (RMSE) and the correlation coefficient (R2). The result shows that models considering generator speed and pitch angle performs better with lowest MAE and RMSE and highest R2 values. A case study illustrated that the AI models, using wind speed, generator speed and pitch angle inputs, would have successfully detected a pitch fault due to the slip ring malfunction nearly 5 hours earlier than the existing fault detection mechanisms.