The method for quantitative prediction of the reservoir connectivity is given for one of the oilfields of Tomskarea. The studied upper Jurassic reservoir was formed in coastal environment and penetrated with a smallnumber of wells. The volume of recoverable oil is depend on chosen method for lithology or N/G modelling. Thecalculations, which were done on geology and simulation models, showed that recoveryg factor changescan reach 35% and related to reservoir connectivity. The Tomsk area reservoirs geomodels, which haveproduction history and total oil production, is higher than 60% of recoverable oil reserves, were studied inorder to develop method to predict connectivity coefficient of reservoirs. The relations between reservoirheterogeneity and sedimentology features and log data were found. It allowed making prediction forconnectivity coefficient value for case study reservoir. In addition, production data of two wells of case study reservoir were considered. Analysis of wells fromother Tomsk reservoirs with the same completion type shows that connectivity value is related with wellproduction rate drop factor. This regression allows predicting desired connectivity coefficient for case studyreservoir in order to provide appropriate heterogeneity to match model with production history data. Finally, the modeling method with continuous N/G property was chosen for geomodel. This is the onlymethod, which allows achieving heterogeneity level that corresponds to sedimentary environments, logfeatures and production history data. As a result, more accurate geomodel was created which has loweruncertainty lever in prediction of reservoir properties in non-drilled part of the reservoir.