Reliability-based design (RBD) is widely applied to meet the reliability requirement at a reduced cost. Current RBD methods require complete distributions of basic random variables, but in real applications, such as the sealing design for hydraulic piston pumps, only statistical moments may be available. The reliability obtained from the RBD may also be sensitive to possible variations in distribution parameters or moments of basic random variables. To deal with these two challenges in the sealing design for hydraulic piston pumps, this study employs a robust reliability-based design methodology based on statistical moments. The optimization not only minimizes the sealing area of friction pairs, but also the sensitivity of the reliability. As a result, the reliability requirement is maintained with the increased robustness of the reliability. An improved higher moment (up to the forth moment) is also employed for higher accuracy with the associated analytical sensitivity indexes. The robust RBD methodology is successfully applied to the design of the seal of the slipper-swash plate friction pairs for a hydraulic piston pump. The application includes the reliability analysis, reliability sensitivity analysis, and robust RBD. The results demonstrate that robust RBD can produce more robust reliability than a traditional RBD methodology.