Endogenous and exogenous uncertainties exert significant influences on energy planning. In this study, we propose a systematic methodology to excavate the uncertainty space, by combining mix-integer linear programming (MILP), Monte Carlo simulation, and machine learning for quantification of the uncertainty impacts on a national-level energy system from global and local perspectives. This approach allows in-depth correlation analysis highlighting potential synergies and risks in the energy transition, and can be easily applied emissions) and energy autonomy can be achieved by 2050, but the energy system's configuration varies significantly under uncertainty. Through conditional distribution analyses, carbon capture and storage (CCS), Photovoltaic (PV), and wood gasification show the most strong correlation for decarbonization. This study is based on the whole uncertainty space taking into account heterogeneous uncertainties, leading to increased reliability compared to sensitivity analysis from single scenarios' comparisons. The synergy between energy models and artificial intelligence (AI) is promising to be widely applied in energy planning area, particularly for emerging technologies with large uncertainty in development.