Abstract: A new local-information-based segmenting method (LI-RGA) of irregular seismic ev ents is given. This method solves the fussy decision of homogeneity criteria in minimum-variance-based region growing algorithm (R-RGA) and the very slow segmen ting in the assessment-function-based region growing algorithm (RM-RGA). The seg menting precision of LI-RGA is not so good as the RM-RGA, but the precision is e nough to interpret of irregular seismic events. According to the features of sei smic data, the traditional ray-casting algorithm is modified, and the displaying time is saved. By defining segmented regions as a mask, then combining them wit h the advanced ray casting　algorithm to change the transfer function, the irreg ular seismic event can be rendered definitely. The test result proves that this method is more effective than the traditional ray-casting algorithm, and can con struct and display unidentifiable irregular seismic events of data set and other important seismic events.