Background: Given inconsistent results in the literature, our objective was to examine the role of early parental feeding practices in children’s growth. Methods: Analyses were based on 1245 children from the EDEN mother–child cohort. Parental feeding practices were assessed at the 2-year follow-up by using the Comprehensive Feeding Practices Questionnaire. International Obesity Task Force BMI z-scores were derived from weight and height assessed at 2, 4, 6 and 8 years. Associations between parental feeding practices and child BMI z-scores at 4, 6 and 8 years were assessed by multivariable linear regressions, notably adjusted for 2-year BMI z-score. Analyses were stratified by child sex when relevant. Moreover, interaction and mediation analyses were respectively performed to assess whether parental feeding practices could moderate or mediate the associations between early and later growth. Results: For a given BMI z-score at 2 years, parental restriction for weight at 2 years was positively associated with child BMI z-scores from 4 to 8 years (at 8 years: β [95% CI] = 0.09 [0.01; 0.16]). Among boys only, high use of food as a reward was positively associated with later BMI z-scores (at 8 years: β [95% CI] = 0.15 [0.03; 0.27]). Parental feeding practices were not moderating factors in the associations between early and later growth. Parental restriction for weight was a mediating factor in the associations between 2-year BMI z-score and BMI z-scores up to 8 years (mediation: 2.69% [0.27%; 5.11%] of the total effect at 8 years). Conclusions: Restriction for weight reasons, often used by parents in response to the child’s high appetite in infancy, appears to lie on the pathway between early and later BMI, but not restriction for health, suggesting that parental way of restricting the child’s food intake matters.