The present study was conducted to investigate the effect of Blended learning-Based Social Networks on female second-year high school students’ self-efficacy in math in Gha’emshahr. The population of study consisted of 6263 female second-year high school students in Gha’emshahr. They were studying in 42 schools in 93-94 school year. Out of schools with ICT facilities, one school was randomly selected, then two second grade classes were selected as experimental and control groups. Matius et al’s Self-efficacy questionnaire was used to gather required information on self-efficacy. It was used as a pre-test, survey and post-test instrument in both control and experimental group. First pre-test was administered. Then, during 8 weeks of first term of school year, educational contents were taught using social networks and traditional methods for experimental and control groups, respectively. After training, both groups were assessed in terms of their self-efficacy. Survey test was administered two months after training. Using frequent assessments, the results revealed that there was a significance difference in the self-efficacy of experimental and control groups (P≥./001). Experimental group continued their superiority in self-efficacy two months after training. Research results demonstrated the positive effect of Blended learning-Based Social Networks on self-efficacy.