<p>Droughts are regarded as multifaceted and harmful natural risks, with an extensive socio-economic and environmental effects. Since various factors influence the occurrence of drought it is challenging to accurately monitor and predict the occurrence of droughts. Further, the challenge is intensified due to climate change as it increases the frequency, severity, duration and spatial extent of droughts in recent decades. Global climate models (GCMs), providing hydroclimatic variations under different changing climate scenarios, is a plausible data source to examine the droughts. However, limited spatial resolution and associated systematic bias pose challenges in adopting the hydroclimate variable obtained from GCMs with our post-processing. This research is a critical review of current development in the field of drought studies with special emphasis laid on the use of GCMs in drought monitoring and prediction. This study presents various statistical and dynamical drought prediction methods, bias correction methods, downscaling methods and Multi-Model Ensemble (MME) frameworks to enhance the accuracy of GCM-based drought measurements. Special attention is paid to the emergence of differential-weighted MME methods and the increased importance of machine learning tools in the improvement of bias correction, down-scaling, and ensemble weighting. The review finds out that there are still gaps that exist concerning biases, regional scale model behavior, and physical inadequate representation of major climate processes in GCMs. Lastly, it provides the future research direction, which presents the necessity to establish combined frameworks that integrate better climate process models, observational observations, ensemble models, and machine learning for precise drought assessment and predictions.</p>