The most appropriate method to predict postoperative outcomes in patients with severe obesity undergoing elective non-bariatric surgery is not known. We conducted a single-centre prospective cohort study in patients with a body mass index of at least 35 kg/m2 undergoing non-bariatric, non-cardiac surgery. Patients completed the six-minute walk test prior to surgery. Disability was measured before and six months after surgery using the World Health Organization disability assessment schedule. Primary outcome measures included quality of recovery at one month and significantly increased disability at six months after surgery. A total of 293 patients participated. The median body mass index was 41 kg/m2. Patients generally recovered well. Following surgery, 9% of patients experienced a poor quality of recovery and 7% developed a significant increase in disability. The proportion of patients free from clinically significant disability increased from 66% prior to surgery to 90% at six months after surgery. The distance walked in six minutes was weakly predictive of poor recovery at one month and significantly increased disability at six months; weight and body mass index were not. The area under the receiver operating characteristic curve was 0.65 (95% confidence intervals 0.51 to 0.78) for poor recovery and 0.64 (95% confidence intervals 0.51 to 0.77) for increased disability. A preoperative six-minute walk test distance of 308 m was the best cut-off value for predicting increased postoperative disability (sensitivity 0.68, specificity 0.63). The six-minute walk test was most discriminatory at shorter distances. This population of patients with severe obesity appeared to recover well and had few adverse outcomes. The degree of functional capacity was more important than the degree of obesity in predicting postoperative outcomes.