The purpose of this study was to identify the longitudinal causality between gambling beliefs and gambling behavior. An online survey was conducted 3 times across 10 months (January, May, and October) among adult gamblers (N=340) who had more than one year of regular gambling behavior, at least once a month (N=340, 64.7% males). The mean age of the subjects was 40.34 y (SD=0.43). The causality between irrational gambling beliefs and gambling behavior was analysed using autoregressive cross-lagged modeling. Analysis showed that gambling beliefs affected gambling behavior after 5 months (B=0.152, p<0.001), and gambling behavior also affected gambling beliefs after 5 months (B=0.090, p<0.01). These results explain that there is a mutual causality between irrational gambling beliefs and gambling behavior. Thus, the gambling disorder treatment programs should take into account not only cognitive behavioral theory but the cognitive dissonance perspective.