We examined psychometric properties of a Modified Physical Activity Questionnaire for Children (MPAQ-C). Thirty-two parents (Study 1), 40 students (6–9 years) and one of each student’s parents (Study 2), and 625 parents (Study 3) completed the MPAQ-C. The MPAQ-C (six items) measured children’s physical activity (PA) after school, and during evenings and weekends for 7 days. Test–retest reliability (Study 1) and convergent validity (Study 2) were measured. Factor validity of the MPAQ-C (Study 3) was examined using confirmatory factor analysis. A single-factor model of the MPAQ-C fit the data well (χ2(9) = 42.78, p < .001; comparative fit index[CFI] = .977; non-normed fit index [NNFI] = .962; root mean square error of approximation [RMSEA] = .079 [90% confidence interval {CI} = .057 to .11]), with good test–retest reliability, composite reliability (.80) and convergent validity. The factor loadings of MPAQ-C were invariant across fathers/mothers (Δχ2(6) = 3.44, p > .05). The MPAQ-C is a suitable parent proxy for measuring young Chinese children’s PA.