In the context of dual carbon goals, the large-scale development of near-zero energy buildings is emerging as a trend. For optimizing energy system configurations, traditional designs typically focus on deterministic conditions. However, given the uncertainties in various factors during actual operation, building loads are uncertain, leading to mismatches between energy capacity and load, resulting in energy waste or failure to meet the buildings' energy demands. This study introduces an optimization method for the configuration of nearly-zero energy building energy systems, considering load uncertainties, by characterizing uncertain parameters and developing scenario sets, aiming at minimizing total lifecycle costs and carbon emissions, thereby enhancing energy efficiency and system reliability. The results show that compared to the initial configuration, both the deterministic optimization configuration and the configuration considering load uncertainties have enhanced the energy system's performance. Given that different optimization schemes have differing LCC and LCCO2, we evaluate these schemes based on initial investment cost, equipment carbon emissions, and the likelihood of meeting load demand as key performance indicators. The deterministic optimization configuration reduced the system's initial investment cost by 60.99% and equipment carbon emissions by 29%, but it did not increase the load guarantee rate. While the configuration considering load uncertainties reduced initial investment costs by 58.61% and equipment carbon emissions by 15.8% and increased the probability of meeting load dem1ands by 133.69%.