The large-scale integration of renewable energy has significantly reduced power system frequency support capabilities. To enhance grid frequency stability, grid codes in multiple countries now mandate renewable power plants (RPPs) to provide virtual inertia and primary frequency regulation (PFR). However, the confidentiality and lack of standardization surrounding key RPP frequency regulation control parameters hinder grid operators’ ability to accurately assess their actual frequency regulation capabilities—a critical requirement for ensuring system stability. Existing assessment methods often overlook response delays, resulting in capability overestimation. To address this, this paper proposes a frequency regulation capability assessment method for RPPs that explicitly incorporates response delays. Firstly, a plant-level frequency regulation model is established, integrating control delay, power response dynamics, and control strategies. Secondly, cross-correlation analysis rapidly identifies control delay times, enabling construction of a delay-compensated discrete state-space equation that accurately characterizes RPP active power-frequency responses. Furthermore, the Levenberg-Marquardt (LM) optimization algorithm simultaneously identifies key control parameters, including virtual inertia and PFR coefficients. Finally, simulations on the modified IEEE 39-bus system validate the method’s accuracy and effectiveness across multiple scenarios.