Introduction The aim of this study was to systematically review the evidence on the effectiveness of exercise-based telemedicine in chronic pain. Methods We searched the Cochrane, PubMed, MEDLINE, EMBASE, CINAHL and PEDRO databases from 2000 to 2015 for randomised controlled trials, comparing exercise-based telemedicine intervention to no intervention or usual care in adults with chronic pain. Primary outcome data were pooled using random effect meta-analysis. Primary outcomes were pain, physical activity (PA), limitations in activities of daily living (ADL) and quality of life (QoL). Secondary outcomes were barriers, facilitators and usability of telemedicine. Results Sixteen studies were included. Meta-analyses were performed in three subgroups of studies with comparable control conditions. Telemedicine versus no intervention showed significantly lower pain scores (MD -0.57, 95% CI -0.81; -0.34), but not for telemedicine versus usual care (MD -0.08, 95% CI -0.41; 0.26) or in addition to usual care (MD -0.25, 95% CI -1.50; 1.00). Telemedicine compared to no intervention showed non-significant effects for PA (MD 19.93 min/week, 95% CI -5.20; 45.06) and significantly diminished ADL limitations (SMD -0.20, 95% CI -0.29; -0.12). No differences were found for telemedicine in addition to usual care for PA or for ADL (SMD 0.16, 95% CI -0.66; 0.34). Telemedicine versus usual care showed no differences for ADL (SMD 0.08, 95% CI -0.37; 0.53). No differences were found for telemedicine compared to the three control groups for QoL. Limited information was found on the secondary outcomes. Conclusions Exercise-based telemedicine interventions do not seem to have added value to usual care. As substitution of usual care, telemedicine might be applicable but due to limited quality of the evidence, further exploration is needed for the rapidly developing field of telemedicine.