Background <p>Early childhood nutrition is crucial for long-term health; however, assessing dietary intake in young children is challenging due to the reliance on parent/caregiver reporting and the need for culturally relevant tools. The aim of this study (the Polish Infant and Toddler Nutrition Study, PITNUTS 2024) was to address this gap by developing and evaluating the reproducibility of a new 85-item Child Food Frequency Questionnaire (85-item Child-FFQ) and associated diet quality scores for Polish children aged 13–72 months.</p> Methods <p>A cross-sectional, nationally representative sample of 700 Polish children from two age groups (13–36 and 37–72 months) was recruited, and 100 children were randomly selected from each age group for the reproducibility study. The 85-item Child-FFQ was administered twice to the same parent as a proxy reporter. Three diet quality scores were developed: the children’s pro-Healthy Diet Score, the children’s non-Healthy Diet Score, the children’s Ultra-Processed Food Score. Statistical analyses included comparisons of mean values for food consumption frequency and diet quality scores, cross-classification, kappa statistic, intraclass correlation coefficient (ICC), Spearman’s correlation, and Bland-Altman plots.</p> Results <p>The 85-item Child-FFQ showed good test–retest reproducibility. Classification agreement for food consumption frequency ranged from 74% to 97% for children aged 13–36 months, and 77% to 95% for children aged 37–72 months. The kappa statistic indicated moderate to excellent agreement between test and retest of the FFQ. Diet quality scores demonstrated high reproducibility, with classification agreement ranging from 77% to 92%, the kappa statistic ranging from 0.62 to 0.86, and good to excellent correlation coefficients (ICC: 0.886–0.950, Spearman’s r: 0.889–0.947). The Bland-Altman analysis revealed a small negative shift between test and retest of the diet quality scores and moderately wide limits of agreement.</p> Conclusions <p>The newly developed 85-item Child-FFQ is a reliable tool for assessing habitual food consumption frequency and diet quality in Polish children aged 13–72 months. This tool has the potential to help monitor food consumption and diet quality variations over time, though this capacity needs to be validated through future longitudinal studies.</p> Trial registration <p>This trial was registered at www.clinicaltrials.gov (NCT06417151, 16th May 2024).</p>