This paper presents a linear-like adaptive control theory for nonlinear systems. The system to be controlled is first described by a so called quasi-ARMAX model which is a linear-like nonlinear model. The quasi-ARMAX model is distinctive to other models in that its polynomials in backward shift operator (filters) are time-varying, but they can be treated as commutable ones simultaneously. By using this special property of filters effectively, we then develop an adaptive control scheme in a similar way to linear control theory, in which a group of adaptive fuzzy models are used to re-parameterize the unknown coefficients of controller. The effectiveness of the proposed scheme is confirmed through numerical simulations.