Hot extremes and ozone pollution have long been known detrimental to public health, but until very recently disproportionate health impacts from their joint occurrence-compound hot and ozone extremes (CHOEs)-have not been sufficiently aware of. Based on high-quality observations of air temperature and surface ozone concentration, we here examined the features of urban CHOEs and their dependence on city population and background climates. Results show the ozone-temperature slope (m(O3-T)) is significantly correlated with city population size, and the correlation is much weaker (0.38) in moister and cloudier cities (Cluster I) as opposed to that (0.69) in drier and sunnier cities (Cluster II). Larger cities are more susceptible to CHOEs with Cluster II megacities (population > 10 million) registering about 8 days of CHOEs during ozone season but similar to 3 days for small cities (population <= 1 million). Most cities experience elevated risks of CHOEs in urban areas relative to surrounding rural areas, especially so for densely-populated cities and those located in drier and sunnier environments. This study emphasizes the importance and urgency of emission reduction to mitigate the health burden from not only hot extremes but also their hazardous compounding with surface ozone.