This study aimed to investigate the effectiveness of training social problem solving strategies on self-esteem in second-grade guidance schools female students in Yazd. The research design was quasi-experimental with pre-test, post-test, and control group. (Without intervention) and pcychological placebo group (except for intervention of social problem solving strategies). The population in this study comprised of second-grade guidance school female students in Yazd who were enrolled in school year 2012-2013. The sample of this study consisted of 67 students who were selected using multistage random cluster sampling and were assigned in three groups (experimental: N=20, control: N=20, and psychological placebo: N=27). As for self- esteem a pre-test was performed in all the experimental, control and placebo groups. Social problem solving strategies training was performed on experimental group in 10 sessions 60 minutes each. Following the experimental intervention, self-esteem of the three groups were measured again. Cooper Smith Self-esteem Inventory was used for measuring self-esteem. Analysis of covariance and Bonferroni Post hoc test were used to check the results. The results showed that in the experimental group social problem solving strategies training increased adjusted mean scores of self-esteem scores (P<0.001) and mean scores on the subscales of self-esteem, including total self-esteem and educational self-esteem, were increased. But this intervention could not increase adjusted mean scores of family self-esteem and social self-esteem. According to the research findings, social problem solving strategies training increased self-esteem. So this method can be used as one of the possible ways to increase self-esteem in the critical period of puberty.