High-pressure compressor (HPC) blades of aero engines inevitably exhibit various uncertain geometric deviations, which deteriorate engine performance and increase maintenance costs. Although the condition-based maintenance (CBM) strategy is increasingly adopted to reduce costs by tailoring repair actions based on condition monitoring data, maintenance practices often still rely on original equipment manufacturer (OEM) recommendations. To further refine the CBM strategy, this paper proposes an uncertainty quantification method based on the engine performance digital twin (PDT) model to quantify the impact of HPC blade geometric deviations on overall engine performance. The PDT model is developed by coupling computational fluid dynamics simulations with a zero-dimensional performance model using real operating data and is validated for high predictive accuracy. Surrogate models based on support vector regression are employed to efficiently quantify the impact of combined geometric deviations. The results show that combined deviations cause reductions in mass flow, pressure ratio, and efficiency while increasing exhaust gas temperature and specific fuel consumption. The proposed methodology is applied to a CBM scenario to demonstrate its effectiveness. In the real maintenance process, this method enables the prediction of performance after repair, facilitating optimized maintenance strategies.