Substance use disorders have a profound impact on the course of severe mental illnesses and on the family, but little research has evaluated the impact of family intervention for this population. To address this question, a randomized controlled trial was conducted comparing a brief (2-3 mo) Family Education (ED) program with a longer-term (9-18 mo) program that combined education with teaching communication and problem-solving skills, Family Intervention for Dual Disorders (FIDD). A total of 108 clients (77% schizophrenia-spectrum) and a key relative were randomized to either ED or FIDD and assessed at baseline and every 6 months for 3 years. Rates of retention of families in both programs were moderate. Intent-to-treat analyses indicated that clients in both programs improved in psychiatric, substance abuse, and functional outcomes, as did key relatives in knowledge of co-occurring disorders, burden, and mental health functioning. Clients in FIDD had significantly less severe overall psychiatric symptoms and psychotic symptoms and tended to improve more in functioning. Relatives in FIDD improved more in mental health functioning and knowledge of co-occurring disorders. There were no consistent differences between the programs in substance abuse severity or family burden. The findings support the utility of family intervention for co-occurring disorders, and the added benefits of communication and problem-solving training, but also suggest the need to modify these programs to retain more families in treatment in order to provide them with the information and skills they need to overcome the effects of these disorders.