The caseloads of food stamp program and welfare program experienced sharp drops among non-metropolitan single female-headed families with children (SFHFwC) after welfare reform in 1996. There is concern that leaving welfare program has led to nonparticipation in food stamp program due to multiple reasons. This study conducts an empirical analysis of non-metropolitan SFHFwC's participation in food stamp program and welfare program, using the 1993 and 1999 CPS data. Results from using reduced-form probit model, bivariate probit model and structural probit model are compared. The reduced-form probit model considers participation in both programs as functions of all exogenous variables. The bivariate probit model allows correlation between the participation in the two programs. The structural probit model considers participation in welfare as an endogenous variable of FSP participation. Empirical analysis is also conducted through the comparison across years and between southern and the remaining U.S. states. We found that welfare participation has a significant and positive impact on FSP participation. Households leaving welfare are more likely not to participate in FSP. Further, welfare participation is more important in determining FSP participation in 1999 than in 1993. Residence in South has a significant and negative impact in welfare participation in 1999, but has no significant impact on FSP participation in either year. Incomes and number of young children are the other major determinants of participation in both programs. Policy suggestions include improving economic situation to reduce FSP caseloads on one side, and providing more information about FSP eligibility to welfare leavers to ensure their food safety on the other side. / Master of Science