Aims/hypothesis - The prospective association between cardiorespiratory fitness (CRF) measured in young adulthood and middle age on development of prediabetes, defined as impaired fasting glucose and/or impaired glucose tolerance, or diabetes by middle age remains unknown. We hypothesised that higher fitness levels would be associated with reduced risk for developing incident prediabetes/diabetes by middle age. Methods - Participants were from the Coronary Artery Risk Development in Young Adults (CARDIA) study who were free from prediabetes/diabetes at baseline (year 0 [Y0]: 1985–1986). CRF was quantified by treadmill duration (converted to metabolic equivalents [METs]) at Y0, Y7 and Y20 and prediabetes/diabetes status was assessed at Y0, Y7, Y10, Y15, Y20 and Y25. We use an extended Cox model with CRF as the primary time-varying exposure. BMI was included as a time-varying covariate. The outcome was development of either prediabetes or diabetes after Y0. Model 1 included age, race, sex, field centre, CRF and BMI. Model 2 additionally included baseline (Y0) smoking, energy intake, alcohol intake, education, systolic BP, BP medication use and lipid profile. Results - Higher fitness was associated with lower risk for developing incident prediabetes/diabetes (difference of 1 MET: HR 0.99898 [95% CI 0.99861, 0.99940], p < 0.01), which persisted (difference of 1 MET: HR 0.99872 [95% CI 0.99840, 0.99904], p < 0.01] when adjusting for covariates. Conclusions/interpretation - Examining participants who had fitness measured from young adulthood to middle age, we found that fitness was associated with lower risk for developing prediabetes/diabetes, even when adjusting for BMI over this time period. These findings emphasise the importance of fitness in reducing the health burden of prediabetes and diabetes.