This study investigates production risk. A multistage stratified random sampling technique was adopted to select sampling unit. In between Cobb Douglas and Linear quadratic model, the linear quadratic model had been picked through feasible generalized least square method. The numerical model, we utilize the information from rice cultivating in Bangladesh. The results show that uneven socioeconomic and farm-specific inputs are creating risk in rice production. Input variables such as area, labour, and fertilizer and managerial factors, for example, experience, schooling, contact with extension, training, natural calamity, member and status indicated a significant impact on rice productions uncertainty. This indicated that both input and managerial factors were important for the rice production.