Previous studies have presented evidence revealing that the dispositional personalities of consumers, or the traits associated with their gambling behaviors and risk preferences, differ across various gambling types. Those studies leave open the question of whether and how personality differences affect domain-specific gambling preferences. Using latent class analysis (LCA), we explored the latent classes of risk preference for different gambling types and the effects of personality traits on these classes. A total of 732 Macau residents completed a questionnaire assessing 13 gambling types and personality traits. Preferences for gambling shown by the participants varied between three latent classes of games: chance gambling, entertainment gambling, and technical gambling. Moreover, not all personality traits consistently predicted preferences for these gambling classes. For instance, aggression-hostility positively predicted a general preference for both chance gambling and technical gambling, but impulsive sensation-seeking positively predicted only a preference for chance gambling. However, neuroticism-anxiety traits showed no such predictive effect for any of the three gambling classes. This study suggests that personality patterns emerge for overall and specific gambling types and underscores the value of LCA to help resolve the longstanding debate over whether gambling is domain specific.