This systematic review and meta-analysis examines the effectiveness of combined diet (not limited to caloric restriction) and exercise interventions, and diet interventions alone to improve physical function in community-dwelling older adults. Randomized clinical trials and observational population-based studies of community-dwelling older adults were selected through comprehensive bibliographic searches in Medline (up to September 2014). Included trials had to assess performance-based measures of physical function such as strength, balance, mobility and gait, and diet measured as diet indexes or food intake. Seven studies were included. Meta-analysis was performed with the inverse variance method under the random effects models. Combined exercise and diet interventions, when compared with control or diet interventions alone, were shown to improve walking speed and performance on the Short Physical Performance Battery (SPPB), although SPPB results failed to show clinical significance. No consistent effect was observed for balance outcomes. Although exercise interventions are known to improve physical function outcomes, based on current data, it is not possible to affirm that a combination of diet and exercise interventions can further improve physical function. The evidence comparing different patterns of diet is scarce, and it is not possible to pinpoint which diet intervention is the most effective.