<p>Drought forecasting and evaluation is essential since it has a detrimental impact on both humans and wildlife. The multi-model ensemble (MME) of General Circulation Models (GCMs) has wide range of applications for the assessment of future drought events. These models offer in-depth studies and forecasts to mitigate the negative impacts on ecosystems and human communities. The application of a multi-model ensemble is employed in the bias correction process for these GCMs at the regional scale. This study aims to introduce a novel Index- Bayesian Model Averaging with Diminishing Outliers (BMADO) for statistical downscaling and mitigating the influence of adverse outcomes. This is suggested using the two-phase ensemble weighting scheme based on Bayesian model averaging (BMA) to weight each GCM. This study uses data from 18 GCM to evaluate the effectiveness of the suggested ensemble weighting technique. The study yielded two key findings. First, our analysis demonstrated that the suggested weighting approach outperformed previous ensemble weighting techniques. Consequently, it promotes the use of new weighting scheme in the ensemble of GCMs for future scenarios. Secondly, it suggests that the prediction of future drought features indicates that droughts will occur more frequently in the Tibet Plateau (TP) region. The study employs three widely recognized performance metrics Normalized Root Mean Square Error (NRMSE), Relative Absolute Error (RAE), and Nash–Sutcliffe Efficiency (NSE) to rigorously evaluate and compare the performance of the proposed weighting scheme against the Equal Weighted Average (EWA) approach. Validation of the study evaluates that the proposed method’s average NRMSE (0.2587), outperforming EWA’s 0.2668. Overall, the research enhances the precision of future drought characterization through refinement of the multi-model ensemble of GCMs simulations.</p>