Objective To compare the efficacy of different types of exercises on global cognition in adults with mild cognitive impairment. Methods We systematically searched six databases from their inception to December 14, 2018, for randomized controlled trials evaluating the efficacy of exercise on global cognition in adults with mild cognitive impairment. For the included studies, the mean changes in global cognition from baseline to endpoint were analyzed. A random-effects model was used to conduct pairwise meta-analyses to directly examine the efficacy of different types of exercises. The consistency model was used to conduct a network meta-analysis to evaluate the relative efficacy and rank probability of different types of exercises. Results Pairwise meta-analyses suggested that there was no significant heterogeneity in the exercises. Aerobic exercises (n=5), resistance exercises (n=4), and mind–body exercises (n=7) all had significant benefits compared to the control. Exergame showed a positive trend for higher efficacy than control (n=1) and mind–body exercise (n=1). The results of the network meta-analysis indicated that the four exercise types all had significant benefits compared to the control, and resistance exercises outperformed mind–body exercises. Resistance exercises were most likely to be ranked 1, followed by exergames, aerobic exercises, and mind–body exercises. Conclusions High-intensity and frequent resistance exercises may be the most effective exercise type to improve global cognition in adults with mild cognitive impairment. More multi-arm designs and randomized controlled trials examining the efficacy of exergames are needed.