Background Research suggests that muscle power is a more critical determinant of physical functioning in older adults than muscle strength. The objective of this study was to systematically review the literature on the effect of power training compared to strength training in older adults on tests for muscle power, two groups of activity-based tests under controlled conditions: generic tests and tests with an emphasis on movement speed, and finally, physical activity level in daily life. Methods A systematic search for randomized controlled trials comparing effects of power training to strength training in older adults was performed in PubMed, Embase, Ebsco/CINAHL, Ebsco/SPORTDiscus, Wiley/Cochrane Library and Scopus. Risk of bias was assessed using the Cochrane Collaboration Tool, and quality of evidence was evaluated using GRADEpro Guideline Development Tool. Standardized mean differenences (SMD) and 95% confidence intervals (CI) were calculated for outcomes separately using a random effects model. Results Fifteen trials and 583 participants were included in the meta-analysis. Results indicated a statistically significant benefit of power training on all reported outcomes (muscle power SMD: 0.99, 95% CI: 0.54 to 1.44, p < 0.001; generic activity-based tests SMD: 0.37, 95% CI 0.06 to 0.68; p = 0.02, activity-based tests emphasizing movement speed SMD: 0.43, 95% CI 0.23 to 0.62, p < 0.001). None of the included studies used physical activity level in daily life as outcome. Conclusions Power training offers more potential for improving muscle power and performance on activity tests in older adults compared to strength training. Future research should assess exercise parameters for power training in older adults. In addition, the validity and reliability of the tests used must be evaluated to establish a standardized test protocol. This protocol should also include measurements of physical activity in daily life.