Background With the increasing health care burden of cancer, public health organizations are increasingly emphasizing the importance of calling people to engage in long-term prevention and periodical detection. How to best deliver behavioral recommendations and health outcomes in messaging is an important issue. Objective This study aims to disaggregate the effects of gain-framed and loss-framed messages on cancer prevention and detection behaviors and intentions and attitudes, which has the potential to inform cancer control programs. Methods A search of three electronic databases (Web of Science, Scopus, and PubMed) was conducted for studies published between January 2000 and December 2020. After a good agreement achieved on a sample by two authors, the article selection (κ=0.8356), quality assessment (κ=0.8137), and data extraction (κ=0.9804) were mainly performed by one author. The standardized mean difference (attitude and intention) and the odds ratio (behaviors) were calculated to evaluate the effectiveness of message framing (gain-framed message and loss-framed message). Calculations were conducted, and figures were produced by Review Manager 5.3. Results The title and abstract of 168 unique citations were scanned, of which 53 were included for a full-text review. A total of 24 randomized controlled trials were included, predominantly examining message framing on cancer prevention and detection behavior change interventions. There were 9 studies that used attitude to predict message framing effect and 16 studies that used intention, whereas 6 studies used behavior to examine the message framing effect directly. The use of loss-framed messages improved cancer detection behavior (OR 0.76, 95% CI 0.64-0.90; P=.001), and the results from subgroup analysis indicated that the effect would be weak with time. No effect of framing was found when effectiveness was assessed by attitudes (prevention: SMD=0.02, 95% CI –0.13 to 0.17; P=.79; detection: SMD=–0.05, 95% CI –0.15 to 0.05; P=.32) or intentions (prevention: SMD=–0.05, 95% CI –0.19 to 0.09; P=.48; detection: SMD=0.02, 95% CI –0.26 to 0.29; P=.92) among studies encouraging cancer prevention and cancer detection. Conclusions Research has shown that it is almost impossible to change people's attitudes or intentions about cancer prevention and detection with a gain-framed or loss-framed message. However, loss-framed messages have achieved preliminary success in persuading people to adopt cancer detection behaviors. Future studies could improve the intervention design to achieve better intervention effectiveness.