Whereas mentoring programs are well received as support services, very little empirical research has been conducted to assess the effectiveness of these programs to meet the diverse needs of different special populations of youth. Potentially useful theoretical orientations (attachment, parental acceptance-rejection, social support, adult development, host provocation) and a sociomotivational model of mentoring are presented to complement Rhodes's (2002) model. Mentoring research literatures for five special populations of youth (abused and neglected youth, youth who have disabilities, pregnant and parenting adolescents, juvenile offenders, academically at-risk students) are critiqued. Systemic, longitudinal research must address the cooccurrence of risk factors, populations, and interventions. We conclude with specific recommendations for future research.