The efficacy of a problem-solving skills training group was evaluated with a sample of mentally ill chemical abusers. Twenty-nine day treatment patients were recruited into the study, and then randomly assigned to either a problem-solving group (i.e., treatment) or to the standard day treatment program (i.e., control). The treatment group received 12 sessions of structured problem-solving training in addition to the standard treatment. Problem-solving skills and perceiced stress were assessed pre- and post-treatment, and at a one-month follow-up. Results indicated that treated patients' problem-solving skills were not enhanced, nor were stress levels substantially reduced at either post-treatment or follow-up. The implications of these findings, especially with respect to research with the mentally ill chemical abuser population, are discussed. © 1990 Association for Advancement of Behavior Therapy.