In this paper, a simple approximate method that converts the distribution of kinematic rupture parameters on a fault into that of physical or dynamic parameters, is proposed. The method in which the static stress drop assumed to be equal to the dynamic stress drop, is applied to investigate rupture dynamics of the 1979 Imperial Valley, California, earthquake (M=6.5). The dynamic rupture process is reconstructed by using a propagating crack, which is a process of expanding area of stress release. If we know the amount of stress drop and rupture time at each point on the fault, dynamic rupture process can be simulated. We use the kinematic parameters previously obtained by using waveform inversion techniques. With constraints of the kinematic parameters, the dynamic rupture process of the event can be reconstructed and several dynamic parameters, stress drop and strength excess (the difference between strength and applied initial stress), can be estimated at each point on the fault. By applying this method to the 1979 Imperial Valley earthquake, we found that the rupture initiated at the low strength excess part first, and then expanded toward the north, and ruptured a large-size high stress drop region located in the north part of the fault, and finally decelerated and stopped at the north edge where the strength is high and stress drop is low. These features are well explained by barrier model [Das and Aki, 1977> and are quite similar to those obtained in 1990 by Quin who investigated the dynamic rupture process by the Monte Carlo technique with a large number of trials. Since repeated calculation is unnecessary, our method is very effective and practical compared with his method. ¿ American Geophysical Union 1992