A multi-stage stochastic programming model is formulated to provide suggestions for logistics companies in purchasing and selling truck fleet inventory. To deal with the uncertain nature of several parameters such as productivity, scenario-based nonanticipativity constraints are introduced. Further, this study presents computational results of a specific case to check feasibility of the model, using cplex optimization package in AMPL. To discuss the function of several current policies, more cases are constructed and solved. This study makes corresponding suggestions concerning new energy vehicle incentive according to model solutions. Currently there is little incentive to purchase electric trucks (E-trucks), but under certain policies such as rebate, E-trucks are strongly favored.