Objective To develop a mathematical ability assessment tool suitable for preschool children, and to conduct reliability and validity analysis, so as to provide a scientific and feasible instrument for the early clinical identification of children with mathematical learning difficulties. Methods The framework and items of the scale were constructed and revised based on expert consultations and preliminary surveys, ultimately forming the Preschool Children's Mathematical Learning Ability Scale. A multi-stage sampling method was employed, and a total of 500 children aged 3 - 6 years were included in the study. Reliability and validity analysis were conducted on the sample data. Results The Preschool Children's Mathematics Learning Ability Scale consisted of 24 items across four dimensions. In terms of validity testing, the exploratory factor analysis yielded a KMO value of 0.945, with four dimensions extracted. Confirmatory factor analysis showed that the scale structure model fit well, with χ²/df=3.033, RMSEA=0.064, CFI=0.917, TLI=0.907, and IFI=0.918. The correlation coefficients between each dimension and the total scale ranged from 0.755 to 0.924. Using the arithmetic subtest of the Wechsler Preschool and Primary Scale of Intelligence-Third Edition as a criterion, the correlation coefficient with the total scale was 0.811, and the correlation coefficients with individual dimensions ranged from 0.589 to 0.777. The reliability tests confirmed that the Cronbach's α coefficient for the total scale was 0.937, with each dimension ranging from 0.787 to 0.872. The split-half reliability of the total scale was 0.911, with each dimension ranging from 0.679 to 0.899. The test-retest reliability of the total scale was 0.986, with each dimensions ranging from 0.822 to 0.927. Conclusion The Preschool Children's Mathematics Learning Ability Scale demonstrates good reliability and validity, providing theoretical basis for the early identification and intervention of children with Mathematics learning difficulties.