An autoregressive (AR) model is presented for isotropic and non-isotropic scattering environments characterized by Rice factor 0 \leq K < \infty. It is shown that at least a second order AR process is required to model the damped sinusoidal autocorrelation function (ACF) inherent in such environments. The model parameters are obtained by minimizing the squared error between the actual and the estimated ACF. This approach is shown to improve the ACF approximation as compared to the traditional Yule-Walker method. The model based simulator exhibits an order of magnitude gain in time complexity as compared to the sum of sinusoids method.