The procedure and main result of a comparative study based on using an autoregressive model and an artificial intelligence technique applied to a Wimax traffic data series forecasting task are presented in this document. The time series forecasting methods being compared are: ANFIS model (Adaptive Network-based Fuzzy Inference Sys-tem) and ARIMA model (Auto-Regressive Integrated Moving Average). This article aims to present significant data showing each technique performance under the criteria of mean square error sum and the required processing time. As a result, in this study ARIMA models developed under RATS platforms are compared to the ANFIS models developed through MATLAB.