This paper is concerned with robust performance analysis of sampled-data systems. In particular, we consider the uncertainty of parameters in the continuous-time plant and study the problem of determining the allowable range of the uncertainty around the origin (or equivalently, the allowable range of the parameter around the nominal value) over which the sampled-data system remains internally stable and a prescribed L-2-induced norm (or equivalently, H-infinity performance) level is retained. We provide an iterative procedure with guaranteed convergence that gives an exact allowable parameter range, together with rigorous arguments for the convergence. This method is free from the assumption that an associated sampled-data system has a compact transfer operator, and thus is readily applicable also to such cases when the uncertain parameter appears in the system dynamics, e.g., in polynomial forms or linear fractional forms.