Although process capability indices (PCIs), which are very useful statistical values to summarize process' dispersion and location, have usability in a wide range, they have some limitations because of crisp definitions of the parameters. Crisp definitions of PCIs cause handicaps for a deep and flexible analysis on process capability analysis (PCA). The fuzzy set theory can be used effectively to add more information and flexibility to PCIs. In this paper the one of the most used PCIs which is called process accuracy index (C(a)) is analyzed based on fuzzy specification limits and fuzzy process mean. The fuzzy formulations of the index C(a) are derived for triangular (TFN) and trapezoidal fuzzy numbers (TrFN).