Abstract: Objective To identify impact factors of dengue fever and provide evidence and support for developing control and prevention strategies. Methods A frequency-matching case-control study was adopted to conduct questionnaire survey and laboratory test on the subjects. Random forest model, single factor analysis and multivariate Logistic regression analysis were carried out to identify the impact factors of dengue fever. Results A total of 294 cases （158 males and 136 females） were included in the case group with an average age of （37.0±13.8） years, and 691 cases （329 males and 362 females） were included in the control group with an average age of （36.6±13.2） years. The first 20 variables selected by the random forest model were introduced into the multivariate Logistic regression model. Multivariate Logistic regression analysis showed that a total of seven factors were independent protecting factors including taking preventive measures outdoors （OR=0.29, 95% CI： 0.20-0.43）, knowing spreading vectors of dengue fever （OR=0.40, 95% CI： 0.25-0.64） and mosquitoes' habitats （OR=0.67, 95% CI： 0.50-0.89）, using air-conditioning in summer （OR=0.60, 95% CI： 0.38-0.90）, using screen windows and doors at home （OR=0.52, 95% CI： 0.34-0.79）, cleaning up the garbage in time （OR=0.43, 95%CI： 0.23-0.79） and working indoors with air-conditioning （OR=0.46, 95% CI： 0.30-0.72）. However, people whose family members （OR=6.94, 95% CI： 2.91-16.56） or friends or colleagues（OR=2.71, 95% CI： 1.56-4.72） suffered from dengue fever were at higher risk for dengue infection. Conclusion The infection of dengue fever is highly associated with individual recognition of the disease, mosquito preventive measures, household environment and working environment. It is essential to take preventive measures targeting these factors to decrease the risk of dengue infection.

Key words: Dengue fever, Impact factor, Case-control study, Random forest model