Objective: In the multivariate risk factor model, Freedy et al. (1992) suggested that three important variables will influence the health of people after a disaster. These are pre-disaster factors (e.g., sex, age), within-disaster factors (e.g., the degree of disaster exposure and subjective feelings about the disaster) and post-disaster factors (e.g., secondary stressors, coping, loss of resources, and socila support). This study examined how these three variables affected the psychological symptoms of people after the Chi-Chi earthquake at two different times. Method: This study collected data from 354 subjects two years after the earthquake and from 222 subjects at a one-year follow-up. Although there was an attrition rate of 37.3%, there was no significant difference between the initial and final subject samples on most background variables. Results: Results of hierarchical regression analysis showed that pre-disaster and within-disaster variables only accounted for little variance on psychological symptoms (psychosomatic complaints, re-experiencing, avoidance and maladaptive adjustment) at time one, while the post-disaster variable accounted for the most variances. At time two, after controlling for psychological symptoms at time one, little variance was contributed by pre-disaster, within-disaster, and time one post-disaster variables; instead the time two post-disaster variable accounted for the most variance. Conclusion: Results from this study suggest that while the disaster itself has immediate effects on psychological health, stresses after the disaster have even more detrimental effect.