Health literacy is a broad and multidimensional construct, making its measurement and conclusions inconsistent. This study aims to compare the patterning of health literacy using different assessment tools and examine their impact on children’s developmental outcomes. A cross-sectional study was conducted with 650 students in Years 7–9 from four secondary schools in Beijing. Health literacy was measured by the eight-item health literacy assessment tool (HLAT, score range 0–37), the six-item Newest Vital Sign (NVS, score range 0–6), and the 16-item Health Literacy Survey (HLS, score range 0–16). Based on Manganello’s health literacy framework, information on upstream factors (e.g., gender, ethnicity, socioeconomic status) and developmental outcomes (e.g., health-promoting behaviours, health service use, global health status) was collected. Overall, the average scores for health literacy were 26.34 ± 5.89, 3.64 ± 1.64, and 13.72 ± 2.94, respectively, for HLAT, NVS, and HLS. The distribution of health literacy varied by socio-demographics and individual characteristics except for gender, no matter which health literacy assessment tool was used. The magnitude of associations between health literacy, its upstream factors and developmental outcomes was greater when using three-domain instruments (HLAT and HLS) than using single-domain instruments (NVS). The approach to health literacy measurement will influence the conclusion. Using multidimensional assessment tools may better capture a child’s health literacy and contribute to the maximum efficiency and effectiveness of school-based health literacy interventions.