Abstract: Unmanned farms have great potential to improve productivity and the quality of agricultural products. However, the investment process must face various uncertainties and potential risks. Based on the basic structure and key technology of unmanned farms, this paper proposes a set of investment risk assessment methods based on Bayesian and Dempster-Shafer Theory, which takes various evaluation indicators as evidence sources and transforms them into trust functions. A total of 14 evaluation indexes were screened to construct the unmanned farms’ investment risk benefit assessment system, and a corresponding database and evaluation system was established, so that investors could choose the evaluation indicators of interest according to their own situation and input the relevant data, and the system would calculate and give the corresponding risk-benefit evaluation report based on the evidence theory. The research results aim to provide a scientific and comprehensive evaluation tool for decision-makers in the field of unmanned farm investment, and promote the sustainable development and innovation of agricultural production..