The purpose of this prospective cohort study was to assess how cardiorespiratory fitness (CRF) of children with probable developmental coordination disorder (DCD) changes over a period of 4.7 years relative to a group of typically developing controls. A school-based sample of children in a large region of Ontario, Canada with 75 out of a possible 92 schools consented to participate. Children enrolled in Grade 4 (mean = 9.9 years, SD = 0.35) at baseline (n = 2278) were followed over the course of 56 months. A total of eight waves of data collection were carried out throughout the study period. The short form of the Bruininks-Oseretsky test of motor proficiency was used to identify children with probable DCD and the maximal speed attained on the Leger 20-m shuttle run to measure CRF. Mixed-effects modeling was used to estimate the change over time in maximal Leger run speed for both groups adjusting for relevant covariates (e.g., gender, BMI, school, activity level, predilection for activity). Children with pDCD had consistently lower maximal run speed relative to controls. The trajectories of run speed in children with probable DCD and those without the disorder differed by gender with pDCD females demonstrating the lowest scores over time. Both genders with probable DCD showed a greater rate of decline in CRF over time relative to the controls. In conclusion, the difference in CRF between children with and without probable DCD is substantial, and it tends to increase over time. This adds to the argument suggesting that interventions intended to improve CRF may be appropriate and necessary for children with motor difficulties. (Contains 2 tables and 1 figure.)