The aim of this study was twofold: (i) to examine the variations of estimated maximal aerobic speed between non, low and high responders and (ii) to analyze the relationships between accumulated training load parameters and variations of maximal aerobic speed in children soccer players. Forty-four male soccer players were assessed three times during the early and mid-season (second to fifth month of the season) and were monitored daily over the period of analysis using the rating of perceived exertion (RPE), recording the training duration (in min) and calculating the session-RPE (sRPE). Pairwise comparisons revealed that maximal aerobic speed (MAS) was greater for the third assessment than the first (p-value [p] = 0.003; standardized effect of Cohen [d] = 0.355) and second (p = 0.013; d = 0.193) assessments. Large correlations were found between MAS and accumulated RPE, accumulated time, and accumulated sRPE. Moreover, non, low and high responders differed in.MAS (p<0.001) with the last group presenting the largest improvement in MAS. Results suggest that children with lower MAS baseline levels will improve more this capacity over the early and mid-season period compared to children with better baseline levels. Moreover, associations between accumulated training load and MAS were found, suggesting that the training effort can be related with aerobic capacity changes.