This paper presents a framework for uncertainty management and treatment, an essential step toward achieving a robust microgrid design. The microgrid is first designed using a deterministic techno-economic optimization model. All uncertain inputs are identified and modeled properly (probabilistic framework for parameters and bootstrap for time series): this is called the Uncertainty Characterization step.An Uncertainty Analysis is then performed using Monte Carlo simulations to quantify the uncertainty impacts on the microgrid performance. Finally, a two-step Global Sensitivity Analysis is performed to identify and rank the uncertain factors that are mainly responsible for microgrid performance variations [1]. Morris’ method is used to discard unimportant factors with low computational cost. Sobol's variance-based method is applied to quantify the contribution of each uncertain input to the output variance.The case study results indicate that the proposed methodology provides a better understanding and handling of uncertainties during the microgrid design stage and acts as a decision-support tool when evaluating microgrids sizing decisions.