Background. Physical training, regardless of the presence of concurrent weight loss, provides numerous health benefits for individuals who are overweight and obese and have or are at risk for cardiovascular disease. Purpose. The purpose of this review was to identify different types of physical training programs (aerobic, resistance, or combined), with or without counseling/diet modifications, and their impact on physical fitness in individuals who have class II and III obesity. Data Sources, Medline and Medline In-Process, EMBASE, the Cochrane Central Register of Controlled Trials, the Cochrane Database of Systematic Reviews, Web of Science, LILACS, CINAHL, SPORTDiscus, PEDro, and PubMed were searched up to June 2017. Study Selection. This review had the following inclusion criteria: body mass index of >= 35 kg/m(2) and age 18 years or older; supervised physical training program; randomized controlled trial; physical fitness outcome (muscular strength, muscular endurance, cardiovascular endurance, and/or flexibility); in English or Portuguese; and available full-text article. Data Extraction. Three reviewers independently extracted data, assessed study risk of bias using the Cochrane tool, and discussed disagreements until consensus was reached. Data Synthesis. Of the 9460 identified articles, 26 were included and 8 were used in a meta-analysis. The meta-analysis showed improvements in walking speed and maximal oxygen uptake but not knee extension strength in the intervention groups. The Cochrane risk-of-bias score indicated that the majority of the data were from randomized controlled trials with a low or unclear risk of bias. Limitations. The large variability of outcomes and interventions made comparisons difficult. Conclusions. A combination of aerobic exercise and resistance exercise, in addition to diet modifications, may improve cardiovascular and muscular endurance in individuals with class II and III obesity. However, conclusions must be interpreted with caution because of the heterogeneity in interventions and outcome measures among the studies and an unclear risk of bias in several studies.