Many authors are interested in the effects that a specific exercise program could have on sagittal spinal curvatures. The purpose of this study was to determine the effects of different exercise programs on thoracic kyphosis and lumbar lordotic angle. This meta-analysis adhered to the PRISMA guideline and it was registered at PROSPERO. Five electronic databases (Pub Med, Cochrane, WOS, PEDro and EBSCO) were searched up to 31 July 2018. Eligible studies were randomized controlled trials that applied an exercise intervention and measured a kyphosis and/or lordotic angle. Study quality was performance by PEDro score. Risk of bias was assessed using the SIGN 50 checklist for randomized controlled trials. External validity was assessed using the EVAT. Ten randomized controlled trials were included for systematic review and meta-analysis. Meta-analysis with a random effect model was performed to infer the pooled estimated standardized mean difference. All studies were RCTs and they involved a total of 284 cases and 255 controls. Seven studies measured kyphosis angle. A large significant effect of the exercise on kyphosis was identified (SMD = -1.400 (95% CI-2.150 a -0.660), p = 0.000). Four studies assessed lordotic angle and moderate but not significant improvement was shown (SMD = -0.530 (95% CI-1.760 a -0.700), p = 0.401). The results suggest that exercise programs may have a positive effect on thoracic kyphosis angle, but no clear effect on lordotic angle. This systematic review suggests that strengthening rather than stretching could be more relevant for kyphosis and both qualities are important for lordosis. It is necessary to conduct more randomized controlled trials to assess the effects of strengthening and/or stretching program on kyphosis and lordotic angle and to establish the type of the exercise that is better for maintaining the sagittal disposition within normal ranges.