Based on the data from 2017 Financial Household Survey, this paper presents an empirical analysis of questionnaire data from 5,390 farming households, using logit and probit binary choice models to derive the factors that significantly affect farming households' loan default. The study found that the factors significantly affected farmers' loan default behavior including internet, credit cards, phone type, smart phone, online shopping, father's education, debt, trust and attention to financial information. Other factors including mother's education, parents' political status, farm income, happiness, choice of return, risk of investment project, total household assets, total household income and total household consumption have no significant impact on farmers' loan default. Among these factors, total farm income, trust and family happiness were found to be passive factors and the remaining variables were found to be active factors. Besides, household network infrastructure including internet, credit card, phone type, smart phone and online shopping were found to have a greater impact on loan default based on marginal effects. Through the analysis of factors that affect loan default, we can better propose corresponding measures to deal with the frequent default of farmers, which is conducive to understanding farmers' loan needs and improving the quality of banks' loans.