Abstarct: Accurate knowledge from reservoir petrophysical characteristics is necessary to predict the future performance of the oil field. One of the significant petrophysical properties of reservoir is porosity. Therefore, improved modeling techniques are necessary to increase the accuracy in the porosity estimation. In the known approach, used different variables for porosity conditional modeling. For example, we can model the lithology and also we can model the porosity in the each block according to estimated lithology. In this thesis, for importance of porosity in the production potential, it has been tried to estimate the effective porosity from the conventional well logs with the help of intelligent methods. Two methods have been used for effective porosity estimation in seven wells of Asmari reservoir in the F oil field. 1-multi laxyer perceptron neural network 2- combination of MLP and conditional probabilities In the first step, effective variables in porosity estimation were determined using multivariate analysis and Bayesian network. After ward mentioned methods have been used for selected variables. In both methods Trainlm better than other optimization algorithm in MATLAB for backpropagation algorithm, and the best result is obtained from multilxayer perceptron with 15 and 20 neurons for first method and 15 and 15 neurons for secondary method in the hidden laxyer. Regression coefficient and MSE of validation data for the MLP and selected variable are equal to 0.947 and 0.016. regression coefficient and MSE of validation data for the conditional method are equal to 0.876 and 0.0403. Finally the 3D porosity model has been prepared using effective porosity information in seven oil wells. According to this model, porosity value in the central and southeast erea of field is more than the other area.