The purpose of this study was to assess whether young children could reliability regulate exercise intensity production after several practice trials, without reference to objective feedback measures. The study used a new 10-point scale (Cart and Load Effort Rating [CALER] Scale), which depicts a child on a bicycle, at various stages of exertion, towing a cart in which the load increases progressively. After warm-up, 20 children, aged 7–10 years, performed an intermittent, effort production protocol at CALER 2, 5, and 8 on a cycle ergometer. This was repeated on three further occasions in the next 4 weeks. An increase in PO across trials (44, 65, and 79 W at CALER 2, 5, and 8, respectively) confirmed that the children understood the scale. A Bland and Altman limits of agreement (LoA) analysis and an intraclass correlation analysis (ICC) between trials (T) indicated that reliability improved with practice. Intertrial comparisons of overall reliability from T1 to T2 and from T3 to T4 ranged from 0.76 to 0.97 and an improvement in the overall bias ± 95% limits of agreement from −12 ± 19 W to 0 ± 10 W. This study is the first to apply more than two repeated effort production trials in young children and provides strong evidence that practice improves the reliability of effort perception in children. The data also provide preliminary evidence for the validity of the CALER Scale in children aged 7–10 years.