Objective: This study evaluated changes in self-efficacy in non-physically active adolescent girls (13-19 years old) who participated in a six-month, empowerment-based exercise intervention programme (EIP). Design: The study used a pre- and post-test randomized group design and included one pre- and one post-test (at six months) and non-physically active adolescent girls (N = 110) were assigned to an intervention group (n = 54) or a comparison group (n = 56). Setting: Two upper secondary schools and five secondary schools, located in the low socio-economic areas of two communities in southern Sweden were involved in the study. Method: The Swedish version of a 10-item General Self-efficacy Scale (GSES) and the Social Barriers to Exercise Self-efficacy Questionnaire (SPBESQ) were used. In addition, BMI and results from a physical fitness test were measured. For statistical analysis, the Mann-Whitney U-test and the Wilcoxon's matched-pairs signed-rank test were used. Results: Analysis showed a statistically significant difference in GSES scores (p = 0.037) between the groups after the EIP was implemented. Girls in the intervention group had increased their levels of general perceived self-efficacy (p = 0.004). Both groups increased their level of physical fitness (intervention, p = 0.06 and control, p = 0.013). BMI increased in the control group (p = 0.031). Conclusions: The EIP had an impact on adolescent girls' general perceived self-efficacy and can be regarded as an outcome of empowerment that indicates the development of the adolescent girls' ability to effectively deal with a variety of stressful situations in general. (Contains 3 tables and 1 figure.)