Objective To investiagte the influence of different starting time of intravenous urokinase after onset of acute myocardial infarction on the efficacy of thrombolysis. Methods One thousand one hundred and thirty eight cases of acute myocardial infarction admitted in 37 collaborative hospitals were given intravenous urokinase with a standard program. Patency of infarct-related coronary artery was assessed by uniform clinical criteria. Patency rate and 4-week mortality in groups of different starting time of urokinase administration after acute onset were compared. Results According to the starting time of urokinase administration after acute onset, the 1138 cases were divided into 4 groups: 128 cases were within 2 hours (h), 461 within 2-4h, 434 within 4-6h and 115 within 6-12h. The patency rates were 71.9%, 70.1%, 63.6% and 40.0%, respectively and 4-week mortalities were 7.0%, 6.5%, 12.2% and 13.9%, respectively in the four groups. The differences of patency rates and 4-week mortalities between the groups within 2 hours and within 2-4 hours were not statistically significant, so these two groups were combined into one, i.e. within 4 hours. The patency rate of the group within 4 hours (70.5%) was significantly higher than those of the groups within 4-6 hours (P < 0.025) and within 6-12 hours (P < 0.001); the 4-week mortality of the former (6.6%) was obviously lower than those of the latter two (all P < 0.01). The patency rate of the group within 4-6 hours was higher (P < 0.001), than that of the group within 6-12 hours but the 4-week mortalities were not statistically different between the two groups. Conclusion Intravenous urokinase starting within 4 hours after onset of acute myocardial infarction is of the best efficacy.