Bangladesh’s agriculture sector is particularly vulnerable due to its sensitivity to flooding. Thus, Bangladeshi farmers always suffer with flood risk and faces substantial crops losses because of lack of crop insurance. The relationship between the flood risk behavior of farmers specifically their risk perception and attitudes and propensity to purchase crop insurance as a risk management strategy in developing countries has not been thoroughly studied in previous research. This study aims to fill this knowledge gap by examining survey responses, personal risk assessments, and experimentally derived risk preferences from 480 agricultural households in flood affected northern Bangladesh. For evaluating farmers’ risk behavior towards flood, the ELCE (equally likely certainty equivalent) method and risk matrix were used in this research. Afterwards, a probit regression model was constructed using these metrics as independent variables. The findings indicate that farmers’ WTP (willingness to pay) decisions were positively correlated with their experimental measures of risk aversion and subjective risk perception behavior. Farmers who were risk averse, as well as those who predicted a higher possibility of flooding, were more likely to obtain crop insurance. Furthermore, among the socio-economic variables farmers’ past flood experience, non-farm earnings, farmers’ group, land ownership, accessibility to extension and informational sources all had a significant impact on their crop insurance purchase decisions. The study findings may assist policymakers in perceiving the flood-risk behavior of farmers, leading to enhanced preventative measures and a more concerted effort to enact flood-reduction legislation. Furthermore, the socio-economic characteristics of rural farm households should be considered carefully when designing a crop insurance policy.