The present study explored the salient characteristics of families experiencing child poverty in Hong Kong by analysing two representative cross-sectional datasets from the 2011 Population Census and 2016 By-census. Specifically, to identify trends in child poverty rates and the risk factors underlying these trends, the relative importance of personal demographic, parental, and household characteristics in predicting the likelihood of experiencing child poverty in 2011 and 2016 was explored with samples of 41,265 and 40,127 children aged under 15 respectively. Analyses indicated that child poverty rates in Hong Kong are 21.8% in 2011 and 22.6% in 2016. Logistic regression analyses revealed that the common use of Putonghua or other Chinese dialects at home, having parents with low human capital and unemployed or employed in a non-high paying position, single-parent family structure, and having a high number of non-working dependents are significantly associated with child poverty between 2011 and 2016. Varying between the two census years, but no less significant in predicting child poverty risk, was the child’s ethnic background. In the light of the findings, the importance of adopting contextually appropriate age ranges of the ‘child’ in child poverty research and some policy implications are discussed.