Energy storage systems (ESSs) serve as flexible resources that significantly contribute to the integration of uncertain renewable energy sources. They can mitigate power fluctuations through fast charging and discharging in ramping services while providing frequency regulation reserves to maintain system stability. However, traditional methods often neglect the uncertainty of storage capacity and the coupling between ramping and frequency regulation constraints, leading to potential capacity violations and inefficient joint optimization. In this paper, we propose a coordinated ramping-frequency regulation optimization strategy considering ESS capacity uncertainty. A probabilistic confidence-based model was developed to characterize the stochastic nature of capacity, and unified ramping-frequency regulation constraints were formulated for ESSs and thermal units. Nonlinear terms were linearized to derive a tractable mixed-integer quadratic programming (MIQP) model. Case studies verified that the proposed approach effectively accounts for capacity uncertainty, enables shared utilization of ramping and regulation capabilities under unified constraints, and enhances the overall efficiency of system flexibility deployment.