Present study developed a methodology to estimate risk attitudes of the producers operating under production risk. It analyzed risk aversion coefficients of the rice producers in flood prone environment of Bangladesh through econometric estimation of risk attitude distribution parameters incorporating the effect of subjective adjustment of human capital in the model. For this purpose, Generalized Method of Moment approximation of profit function was used to account for the stochastic technology with no restriction on preference structure. The results show that the population is characterized by Arrow-Pratt and downside risk aversion. The farmers were found to be moderately risk averse as indicated by smaller values of Arrow-Pratt risk aversion coefficients and relative risk premium.