IEEE802.16 (WiMAX) is a standard that supports high data rate in a wide area with multi-traffic communication, low implementation and the possibility of creating broadcast, multicast and mesh networks. In this paper the Hidden Markov Model as a Discrete Channel Model has been employed to model the burst errors generated from IEEE 802.16/WiMAX; moreover, the precise Hidden Markov Models using Baum-Welch Algorithm have been obtained by estimating the optimal order of these models with comparing statistics such as Average log-likelihood, Probability of Error, 𝑃(0𝑚|1) and Auto-Correlation function. Additionally, the parameters of the best models have been derived. The impacts of a number of Baum-Welch Algorithm iterations and the modulation order on the optimal order estimation with respect to different (𝑇𝑠) were investigated using extensive simulations.