An approach in solving defect shape identification problems is proposed in this paper. This approach is developed with the boundary element method in conjunction with the Kalman filtering technique, especially the extended Kalman filtering technique. The boundary element approach is utilized by reason of easyness on lemeshing of the defect shape in identification process. Applicability and effectiveness of the proposed method are examined with numerical simulation of several identification examples on a circular or an elliptic defect in 2-D elastostatics.