Background: Already in preschool, very preterm (VP) children perform worse than term born-children on preschool mathematical skills tests. Strong associations have been found between preschool mathematical skills, cognition and visual-motor integration. Aims: To compare VP children and their term-born peers on preschool mathematical achievement at the corrected age of five years, and determine whether cognitive, visual-perceptive, visual-motor, and motor-coordination skills, account for any significant differences observed. Study design: Single-center, consecutive cohort study with a term-born comparison group. Subjects: 54 five-year-old VP children and 28 term-born comparison children. Outcome measures: Standardized test for preschool mathematical skills (Dutch pupil monitoring system), cognitive skills (Wechsler Preschool and Primary Scale for Intelligence - third edition), visual-perception, visual-motor integration, and motor-coordination (Beery Visual-Motor Integration test - sixth edition). Group differences were analyzed with ANCOVAs, adjusting for maternal education, preschool grade, and time of assessment. Sobel's mediation analyses tested for possible mediation effects. Results: Preschool mathematical skills and visual perceptive skills were significantly lower in VP children than in term-born children (Cohen's d = 0.63, p = 0.01; Cohen's d = 0.84, p < 0.01, respectively). Sobel's test indicated a significant mediating effect of visual perceptive skills on the association between VP birth and preschool mathematical skills. Conclusions: At preschool age, VP children have poorer preschool mathematical skills compared to term-born peers; deficits that were fully accounted for by poor visual perceptive skills. Our findings indicate the relevance of screening visual perceptive skills at preschool age, enabling timely identification of children at risk for mathematical difficulties.