Abstract:Generally speaking,the procedure for us to use pattern recognition method to do classification and evaluation of seismic parameters includes three steps:learning,clustering and evaluation. The method is mainly characterized by learning the known class samples to form a membership function,which is then taken as a classifier to evaluate the clustering result of the unknown class samples, thus causing seismic parameters to reflect visual geological informations. The paper describes the method in principle and its procedure. The pattern recognition of real seismic parameters proves it a good seismic parameter interpretation method, which can be considered as an effective means for quantitative seismic interpretation.