Introduction The annual summer peak in childhood asthma hospitalizations after school recommences occurs in both northern and southern hemispheres. This study quantifies the timing and magnitude of the effect of school term and returning to school throughout the year on childhood asthma. Methods We analyzed daily emergency asthma hospital admissions (1994–2003) in three climatically diverse Australian cities (Melbourne, Brisbane and Canberra) by age (0–4 years and 5–14 years). Using Poisson regression by week of school term (1–11) with reference to the preceding vacation period, we controlled for seasons and adjusted for temperature, humidity and long-term trends. Results A return-to-school effect was apparent for all three cities and was not limited to summer. Among school-aged children admissions rose from the first week of Term 1, peaking in week 3 in Melbourne (IRR 4.4, 3.8–5.1) and Brisbane (2.5, 2.1–3.0), and week 4 in Canberra (5.4, 3.7–8.0), returning to baseline by the end of term. Smaller significant peaks were apparent in Terms 2, 3 and 4 in Melbourne, Terms 2 and 4 in Canberra, and Term 4 in Brisbane. The same pattern occurred among younger children: in Term 1 for Melbourne (3.8, 3.4–4.2), Brisbane (3.0, 2.6–3.4) and Canberra (4.5, 3.2–6.3). Terms 2 and 4 showed significant peaks in Melbourne and Canberra. Discussion and Conclusions Childhood asthma emergency hospitalizations increased sharply with recommencement of school, peaked at 3–4 weeks, then declined to vacation levels. The return-to-school effect was strongest in Term 1, supporting the hypothesis that exposure to respiratory viruses after a long vacation (6–7 weeks) through mixing at school can be a significant cause of asthma morbidity. Effects were similar but attenuated in younger children, probably due to viruses brought home by older siblings at school and changed childcare arrangements. Seasonal variation in type and number of viruses circulating in the general population could interact with return-to-school effect, and the different climates in the three cities provide for different seasonal profiles of viruses. Return-to-school had least effect in Term 3 (winter); more virus circulating in the general population could mean children are already exposed with little additional exposure from mixing at school. Stress at the commencement of a new school year is unlikely to contribute much to the Term 1 asthma load, as hospitalizations take several weeks to peak. Quantifying the magnitude and timing of asthma hospitalizations in relation to school term could assist in improved hospital planning and better personal asthma management.