Abstarct: Reservoir characterizations play a critical role in appraising the economic success of reservoir management and development of hydrocarbon reservoir. Porosity and permeability are the most important parameters of the petroleum reservoir, which can be used as inputs to construct perophysical models of the reservoir rock. These parameters can be measured at the laboratory utilizing helium and air injection analysis, respectively. Of course, these methods are time-consuming and expensive for huge amount of problems, which exist in petroleum industries. In recent years, developments of hardware and software programming of computers prove the capabilities of intelligent systems and image analysis. The current study comprises two major parts: Firstly, the petrophysical parameters including porosity and permeability were predicted from petrographic data utilizing intelligent systems. Twelve petrographic parameters were extracted from thin section images and three intelligent systems including neural network, fuzzy logic and neuro-fuzzy were used to predict porosity and permeability. The MSE of the NN, FL and NF methods for prediction of porosity in the test data are 0.0256, 0.0214 and 0.0226, while the MSE of the NN, FL and NF methods for prediction of permeability are 0.0139, 0.0061 and 0.0085. The concept of committee machine is used to improve the accuracy of prediction and the MSEs of CMISs for porosity and permeability are 0.0165 and 0.0056, respectively, which correspond to the R2 values of 0.76 and 0.942, respectively. Comparing the results of this study with core data, shows that intelligent systems were successfully applied in porosity and permeability prediction. Among the intelligent systems, the committee machine with intelligent systems obtained the most accurate results. In the second part of this study, the capabilities of image analysis and pattern recognition were used to develop an automate algorithm for pore space separation. The results show that the proposed algorithm has been successful in separating the visual pore spaces in thin section images. The accuracy of algorithm is tested in two ways. Firstly the output is considered the area of each type of pore spaces and secondly, type of pore spaces is considered as output. The results show that the quadratic discriminant function separates the interagranular and biomoldic porosity with MSE of 9% and 1%, respictevely while the linear discriminant function separates the complete moldic and incomplete moldic porosity with MSE of 6% and 9%. Finally, the mahalonobis discriminant function separates the intergranular pore spaces with MSE of 10%. In the second way, the algorithm is classified th