Abstract: The seismic attributes of 43 types are extracted from Wuerhe Wellblock seismic data, of which 5 are selected as basic parameters of the reservoir porosity prediction model using clustering analysis. And the porosities at wells available are used to make a prediction of the reservoir parameters of target zone by multiple regression and nonparametric regression. The results show good effect for its application to the target zone.

Key words: reservoir parameter, prediction, clustering analysis, correlation coefficient, multiple regression, nonparametric regression