Day-to-day change in ambient temperature is associated with acute myocardial infarction (AMI) attacks, but evidence is scarce about the effects of extreme temperatures on the risk of AMI within hours of exposure. This study investigated the hour-level associations between extreme temperatures and AMI occurrence. State-wide data on AMI patients and temperature during winter and summer of 2013-2015 were obtained for Queensland state of Australia. We employed a fixed time-stratified case-crossover analysis to quantify the risk of AMI associated with temperature within 24 h after exposure. Subgroups analyses by age, gender and disease history were also conducted. We observed a very acute effect of cold on men (occurred 9-10 h after exposure), women (19-22 h after exposure), and the elderly (4-20 h after exposure). Cold was associated with elevated AMI risk for men within 9 h (OR = 2.1, 95 % CI: 1.2-3.6), women within 19 h (OR = 2.5, 95 % CI: 1.0-6.0), and the elderly within 4 h (OR: 2.0, 95 % CI: 1.0-4.0). However, elevated risk of AMI associated with heat occurred 15 h later for men (OR: 3.9; 95 % CI: 1.1-13.9) and 23 h later for adults (OR: 4.1, 95 % CI: 1.1-15.4). People never suffered AMI and the elderly with diabetes or hyperlipidaemia were particularly vulnerable to cold. Those that were particularly vulnerable to heat were men never experienced AMI or having hypertension or having hyperlipidaemia as well as women ever suffered AMI. Effects of temperature on AMI risk at sub-daily timescales should be considered to prevent cardiac events.