Abstract Background Joint hypermobility in athletes is associated with increased risk of knee injuries, but its role in relation to shoulder injuries has not been scrutinized. Therefore, our aim was to synthesize the evidence on the association between joint hypermobility and shoulder injuries in athletes. Methods Data sources were MEDLINE, CINAHL, EMBASE, and SPORTDiscus from inception to 27th February 2021. Eligibility criteria were observational studies of athletes (including military personnel), mean age ≥ 16 years, and with a transparent grouping of those with and without joint hypermobility. A broad definition of joint hypermobility as the exposure was accepted (i.e., generalised joint hypermobility (GJH), shoulder joint hypermobility including joint instability). Shoulder injuries included acute and overuse injuries, and self-reported pain was accepted as a proxy for shoulder injuries. The Odds Ratios (OR) for having shoulder injuries in exposed compared with non-exposed athletes were estimated using a random effects meta-analysis. Subgroup analyses were performed to explore the effect of sex, activity type, sports level, study type, risk of bias, and exposure definition. Risk of bias and the overall quality of evidence were assessed using, respectively, the Newcastle–Ottawa Scale and the Grading of Recommendations Assessment, Development and Evaluation (GRADE). Results Among 6207 records, six studies were included with 2335 (range 118–718) participants (34.1% females; athlete mean age 19.9 years). Athletes with joint hypermobility were more likely to have shoulder injuries compared with athletes without joint hypermobility (OR = 3.25, 95% CI 1.64, 6.43, I² = 75.3%; p = 0.001). Exposure definition (GJH, OR = 1.97, 95% CI 1.32, 2.94; shoulder joint hypermobility, OR = 8.23, 95% CI 3.63, 18.66; p = 0.002) and risk of bias (low, OR = 5.25, 95% CI 2.56, 10.8; high, OR = 1.6, 95% CI 0.78, 3.29; p = 0.024) had large impacts on estimates, while the remaining subgroup analyses showed no differences. The overall quality of evidence was low. Conclusion Joint hypermobility in athletes is associated with a threefold higher odds of having shoulder injuries, highlighting the need for prevention strategies in this population. However, due to low quality of evidence, future research will likely change the estimated strength of the association. Protocol registration Open Science Framework registration osf.io/3wrn9.