A two-stage robust optimization configuration strategy for the new energy town is proposed to address the issue of insufficient power supply reliability in integrated energy systems due to the fluctuation and intermittency characteristics during the integration of a high proportion of renewable energy. In the first stage, historical source-load data is utilized to make preliminary decisions on unit capacity configuration, with the objective of minimizing system configuration costs. In the second stage, polyhedral uncertainty sets are employed to describe the uncertainties of source-load, aiming to minimize system operation costs, and power data predictions for the worst-case scenarios are obtained based on the decision outcomes of the first stage. An uncertainty parameter is then introduced to control the conservativeness of the robust optimization configuration scheme. The model is solved using the column and constraint generation (C&CG) algorithm, which iteratively determines unit capacity configuration and converges to the optimal configuration scheme. A case study of the new energy town in Northern China is conducted, and the results verify the effectiveness and feasibility of the proposed strategy and optimization method, demonstrating their capability to enhance the power supply reliability and economy of the new energy town.