Aiming to find out the reason of this outbreak, and provide a reference for the future prevention and treatment.Methods A retrospective cohort study was conducted. Case definition and case searching was carried out while clinical information such as examination, testing and treatment were collected, additional information were obtained by visiting local villagers and interviewing doctors. The relationship among the onset risk, and the time and amount of consumption was analyzed. Furthermore, the remaining wild honey was collected in order to detect the suspicious poison components. Results 12 patients were all from village S, with a critical rate 58% and a fatality rate 25%. Symptoms included abdominal pain (67%), diarrhea (58%), vomiting (58%), dizziness (58%), nausea (50%) and headache (42%). 58% cases suffered from liver, kidney, and heart damage. Epidemic curve showed that exposure was point source model. These 12 patients all had eaten the wild honey, with an attack rate of 100% while nobody was sick who had not. The median incubation period was 21 h (range:5-29 h). There was a negative correlation between consumption and incubation period (r_s=-0.69, P<0.05). Two toxic pollens of Tripterygium and Macleaya were detected in the honey.Conclusion The poisoning outbreak was caused by eating wild honey contaminated by two toxic pollens of Tripterygium and Macleaya. Bee raising should be strictly prohibited at places where poisonous plants grow and people should not eat wild honey at anytime and anywhere.