Muscle tissue typically contains only small amounts of adipose tissue, and the excess deposition of adiposity is considered a pathological phenomenon termed myosteatosis. Several studies have assessed the effects of exercise alone on the severity of myosteatosis, and some studies have reported promising results. We performed a systematic review and meta-analysis to investigate the effects of exercise interventions on myosteatosis (i.e., lipid infiltration and muscle radiation attenuation). Studies were identified through a systematic search of three databases and limited to randomized controlled trials (RCTs) focused on evaluating the effect of exercise interventions on lipid infiltration and/or muscle attenuation in adults. Thirteen studies met the inclusion criteria, and 12 were included in the meta-analysis (n = 465, 84.7% women). The volume of lipid infiltration was decreased in the exercise group compared with the control group [Hedges' g = -0.45, 95% confidence interval (CI), -0.74 to -0.16; P = 0.008, I2 = 0%], and the degree of muscle radiation attenuation was increased (Hedges' g = 0.67, 95% CI, 0.22 to 1.13; P = 0.009, I2 = 59.3%). Based on meta-regression analyses, there were no significant effects of mean age at baseline (lipid infiltration, β = -0.008, 95% CI, -0.035 to 0.019; P = 0.495; muscle radiation attenuation, β = -0.013, 95% CI, -0.036 to 0.011; P = 0.248) and intervention duration (lipid infiltration, β = -0.008, 95% CI, -0.028 to 0.010; P = 0.311; muscle radiation attenuation, β = -0.018, 95% CI, -0.050 to 0.014; P = 0.230) on the effect size estimates. Overall, our findings indicate that exercise can significantly improve muscle quality in populations at risk of developing obesity and sarcopenia-related disability.NEW & NOTEWORTHY Muscle tissue typically contains only small amounts of adipose tissue, and the excess deposition of adiposity is considered a pathological phenomenon termed myosteatosis. The volume of lipid infiltration tissue was decreased in the exercise group compared with the control group, and muscle attenuation coefficient was increased. Based on the meta-regression analyses, the mean age at baseline and intervention duration did not affect the effect size estimates for lipid infiltration tissue and muscle radiation attenuation.