This longitudinal study evaluated the utility of predicting internalizing, externalizing, and comorbid internalizing/externalizing behavior problems in early adolescence from preschool temperament types. Data on 178 individuals (54.5% Male) were obtained from the Helsinki Longitudinal Temperament Project, an ongoing prospective study on the antecedents and sequelae of early childhood temperament in a Finnish birth cohort (N = 6401). Parent ratings on the Finnish version of the Thomas, Chess, and Korn Parent Temperament Questionnaire were used to form seven dichotomous categorical temperament types (predictors) at age 5: Inhibited, Highly Emotional, Impulsive, Typical, Reticent, Passive, and Uninhibited. Gender was also a predictor. Teacher ratings on the Finnish version of the Lambert-Hartsough Adjustment Difficulty Scale were used to form four dichotomous behavior problem categories (outcome categories) at age 12: Internalizing, Externalizing, Comorbid (Internalizing and Externalizing), and Any Behavior Problem (Internalizing or Externalizing). Chi square and logistic regression analyses showed no predictive relationship between any temperament type and any behavior problem outcome category. Male gender strongly predicted Externalizing, Comorbid, and Any Behavior Problem and weakly predicted Internalizing. Boys were two to three times as likely as girls to meet criteria for Any Behavior Problem, Externalizing, and Comorbid behavior problems. Trends in the data suggested differential outcomes for the Highly Emotional and Inhibited types compared to the remaining temperament types; however, small sample size yielded insufficient power to detect possible temperament effects. Implications for temperament theory and further research are discussed.