This systematic review and meta-analysis critically examined the evidence for bodily illusions to modulate pain. Six databases were searched; 2 independent reviewers completed study inclusion, risk of bias assessment, and data extraction. Included studies evaluated the effect of a bodily illusion on pain, comparing results with a control group/condition. Of the 2213 studies identified, 20 studies (21 experiments) were included. Risk of bias was high due to selection bias and lack of blinding. Consistent evidence of pain decrease was found for illusions of the existence of a body part (myoelectric/Sauerbruch prosthesis vs cosmetic/no prosthesis; standardized mean differences = -1.84, 95% CI = -2.67 to -1.00) and 4 to 6 weeks of mirror therapy (standardized mean differences = -1.11, 95% CI = -1.66 to -0.56). Bodily resizing illusions had consistent evidence of pain modulation (in the direction hypothesized). Pooled data found no effect on pain for 1 session of mirror therapy or for incongruent movement illusions (except for comparisons with congruent mirrored movements: incongruent movement illusion significantly increased the odds of experiencing pain). Conflicting results were found for virtual walking illusions (both active and inactive control comparisons). Single studies suggest no effect of resizing illusions on pain evoked by noxious stimuli, no effect of embodiment illusions, but a significant pain decrease with synchronous mirrored stroking in nonresponders to traditional mirror therapy. There is limited evidence to suggest that bodily illusions can alter pain, but some illusions, namely mirror therapy, bodily resizing, and use of functional prostheses show therapeutic promise.