A systematic procedure for designing a robust adaptive control strategy for a class of dynamical systems containing n dominant slow modes and m parasitic fast modes is developed. The result is a corrected adaptive control strategy more robust to parasitic dynamics than conventional methods in the literature that are based on the so-called zeroth-order or uncorrected adaptive control laws. The distinct feature of the proposed controller is its dynamic nature. The key requirement for implementing the dynamic controller is the asymptotic stability of the zero dynamics of the first order corrected reduced model. The results are verified by numerical simulations.