The key to improve the machining quality of workpiece is to decrease the process fluctuation, which requires identifying the fluctuation sources first. For small-batch multistage machining processes of complex aircraft parts, how to identify the fluctuation sources efficiently has become a difficult issue due to the limited shop-floor data and the complicated interactive effects among different stages. Aiming at this issue, a fluctuation evaluation and identification model for small-batch multistage machining process is proposed based on the sensitivity analysis theory. In order to improve the data utilization, an analytical structure of the fluctuation evaluation and identification model for small-batch multistage machining process is presented, which comprises four levels, namely, part level, multistage level, single-stage level and quality feature level. Corresponding to the four levels in the analytical structure, four fluctuation analysis indices are proposed to quantitatively evaluate the fluctuation level of different parts and identify the weak stages and elements that result in the abnormal fluctuation in the process flow. A five-stage deep-hole machining process of aircraft landing gear is used as a case to verify the proposed model.