Abstract Background Smoking is common in people infected with HIV but cessation support is not a routine part of clinical care. The aim was to assess whether smoking is a risk factor for pneumonia in people with HIV and whether smoking cessation ameliorates excess risk. Methods We performed MEDLINE and Embase database searches and included cohort or case-control studies conducted in adult patients infected with HIV extracting a hazard ratio (HR) or odds ratio (OR) that compared the incidence of bacterial pneumonia or pneumonia caused by Pneumocystis jiroveci (PCP) between two smoking categories. Studies were appraised for quality and combined using inverse variance meta-analysis. Results Fourteen cohort and case-control studies were included. Assessment of outcome was good, but assessment of exposure status was poor. Current smokers were at higher risk of bacterial pneumonia than former smokers: HR 1.37 (95% confidence interval (CI): 1.06, 1.78). There was no evidence that former smokers were at higher risk than never smokers: HR 1.24 (95%CI: 0.96, 1.60). Current smokers were at higher risk of bacterial pneumonia than current non-smokers: HR of 1.73 (95%CI: 1.44, 2.06). There was no evidence that smoking increased the incidence of PCP. The HR for current versus non-smokers was 0.94 (95%CI: 0.79, 1.12), but from case-control studies the OR was 1.76 (95%CI: 1.25, 2.48) with heterogeneity. Confined to higher quality studies, the OR was 0.97 (95%CI: 0.81, 1.16). Residual confounding is possible, but available data suggest this is not an adequate explanation. Conclusions Smoking is a risk factor for bacterial pneumonia but not PCP and smoking cessation reduces this risk. See related article: http://www.biomedcentral.com/1741-7015/11/16