BACKGROUND: The use of mobile technologies to prevent STIs is recognised as a promising approach worldwide; however, evidence has been inconclusive, and the field has developed rapidly. With about 1 million new STIs a day globally, up-to-date evidence is urgently needed. OBJECTIVE: To assess the effectiveness of mobile health interventions delivered to participants for preventing STIs and promoting preventive behaviour. METHODS: We searched seven databases and reference lists of 49 related reviews (January 1990-February 2020) and contacted experts in the field. We included randomised controlled trials of mobile interventions delivered to adolescents and adults to prevent sexual transmission of STIs. We conducted meta-analyses and assessed risk of bias and certainty of evidence following Cochrane guidance. RESULTS: After double screening 6683 records, we included 22 trials into the systematic review and 20 into meta-analyses; 18 trials used text messages, 3 used smartphone applications and 1 used Facebook messages as delivery modes. The certainty of evidence regarding intervention effects on STI/HIV occurrence and adverse events was low or very low. There was moderate certainty of evidence that in the short/medium-term text messaging interventions had little or no effect on condom use (standardised mean differences (SMD) 0.02, 95% CI -0.09 to 0.14, nine trials), but increased STI/HIV testing (OR 1.83, 95% CI 1.41 to 2.36, seven trials), although not if the standard-of-care control already contained an active text messaging component (OR 1.00, 95% CI 0.68 to 1.47, two trials). Smartphone application messages also increased STI/HIV testing (risk ratio 1.40, 95% CI 1.22 to 1.60, subgroup analysis, two trials). The effects on other outcomes or of social media or blended interventions is uncertain due to low or very low certainty evidence. CONCLUSIONS: Text messaging interventions probably increase STI/HIV testing but not condom use in the short/medium term. Ongoing trials will report the effects on biological and other outcomes.