Forecasting crude oil price is among the most important issues facing energy economists. Forecasting Suitable of oil price and OPEC crude oil price too, due to involvement many of developing countries in the organization with oil price, can be planning by the organization and its member states, has special significance. Estimation and forecasting of oil price trend is a cumbersome task due to lack of significance historical data and limitations on information regarding of economic indicators affecting the oil price trend. This, in turn, intensifies the amounts of parametric noise, complexity and uncertainty associated with estimation of oil price trend. Nevertheless, the success in formulating a reliable model to describe the complex dynamics of this commodity is limited. In this study, artificial neural network, Neuro-Fuzzy network and wavelet transform-neuro hybrid model and daily OPEC basket oil price, for modeling and prediction of short-term OPEC crude oil price is used. The results of these models based on criteria of measuring forecast accuracy, are compared. Results of the study shows that, firstly, de-noise data can improve network performance and secondly, Neuro-Fuzzy network than other models used in this paper has better predictive power.