Exercise and dietary interventions have been described to positively affect metabolic syndrome (MetS) via molecular-induced changes. The purpose of this study was to investigate the effects of dietary carbohydrate restriction and aerobic exercise (AE) on retinol binding protein 4 (RBP4) and fatty acid binding protein 5 (FABP5) in middle-aged men with MetS. The study had a randomised, double-blinded, parallel-controlled design. Forty middle-aged men with MetS (age: 53·97 ± 2·85 years, BMI = 31·09 ± 1·04 kg/m2) were randomly assigned to four groups, AE (n 10), ketogenic diet (KD; n 10), AE combined with KD (AE + KD; n 10) or control (C; n 10). RBP4, FABP5, body composition (body mass, BMI and body fat), insulin resistance, insulin sensitivity and MetS factors were evaluated prior to and after the 12-week intervention. AE + KD significantly decreased the body fat percentage (P = 0·006), BMI (P = 0·001), Zmets (P = 0·017), RBP4 (P = 0·017) and the homeostasis model of insulin resistance (HOMA-IR) (P = 0·001) as compared with control group and marginally significantly decreased the Zmets as compared with exercise group (P = 0·086). KD significantly decreased RBP4 levels as compared with control group (P = 0·041). Only the AE intervention (P = 0·045) significantly decreased FABP5 levels. Combining intervention of carbohydrate restriction with AE compared with carbohydrate restriction and AE alone improved RBP4, HOMA-IR as well as different body composition and MetS factors in middle-aged men with MetS.