Exercise testing is the gold standard for diagnosing exercise-induced bronchoconstriction in children, but requires considerable cooperation and medical resources. Therefore, fraction of exhaled nitric oxide (FeNO) has been proposed as a tool to predict the need for exercise testing. The objective of this study was to investigate the relationship between FeNO, exercise test results, and a history of respiratory symptoms during exercise in children at risk of asthma. FeNO measurement, exercise testing, and interview about respiratory symptoms during exercise were completed in 224 seven-year-old children from the at-risk Copenhagen Prospective Studies on Asthma in Childhood2000 birth cohort. The associations between FeNO, exercise test results, and reported respiratory symptoms during exercise were analyzed adjusting for gender, respiratory infections, and inhaled corticosteroid treatment. The associations were also analyzed stratified by asthma and atopic status. Of the 224 children, 28 (13%) had an established asthma diagnosis and 58 (26%) had a positive exercise test (≥15% drop in forced expiratory volume in 1 second [FEV1] from baseline). FeNO and bronchial obstruction after exercise were linearly associated with a doubling of FeNO corresponding to a 2.4% drop in FEV1 (95% confidence interval, 0.8-4.1; P < .01). However, a receiver operating characteristic curve analysis showed that the best cutoff of FeNO for predicting exercise test outcome among children who reported respiratory symptoms during exercise was 17 ppb, which only had 74% negative predictive value. There was no association between FeNO and reported respiratory symptoms during exercise (odds ratio = 1.3 [0.8-1.9]; P = .29) or reported symptoms during exercise and exercise test results (odds ratio = 1.0 [1.0-1.1]; P = .12). A history of respiratory symptoms during exercise was not associated with either elevated FeNO or a positive exercise test in children at risk of asthma. FeNO and exercise test results were linearly associated traits, but FeNO could not reliably be used dichotomized to predict the need of exercise testing.