In this paper recent results obtained by the authors for control-oriented identification of stable, discrete-time systems are extended to stable, strictly proper, continuous-time systems. The required a priori information consists of a lower bound on the relative stability, an upper bound on the "steady state" gain, and an upper bound on the rate of "roll-off" of the unknown system. The required experimental information consists of a finite number of possibly corrupt frequency response estimates for the unknown system. Given this information an algorithm is developed which yields both a continuous-time identified model and an explicit bound on the H?? norm of the model error. Like its discrete-time counterpart, this algorithm is a nonlinear function of the data used and is robustly convergent with respect to the a priori system information given.