Abstract: Due to the shortage of resources, project managers are more and more eager to carry out multiple projects in parallel in a limited time and resource range in order to obtain higher benefits. However, there will be a high degree of uncertainty in the parallel development of many projects, which will affect the development of enterprises. The existing project portfolio models all think that the project set can be completed in a fixed planning period, but in the actual project implementation process, the decision makers often adjust the length of the planning period reasonably according to the actual situation, extended the planning period so that the high-yield projects can be included in the project set or shorten the planning period as much as possible to give up the risky projects and release more funds. Therefore, the flexible time horizon is brought into the project portfolio selection model. At the same time, considering that the value of the project will change with time, for the uncertainty of the future value of the project, the binary tree model of real options is used to define the formula of value fluctuation and a new model of project portfolio selection affected by flexible time horizon and value fluctuation is established at the same time, and the traditional net present value method (NPV) is complemented for calculating project value. In the part of case analysis, a set of data is randomly generated, and the rationality and effectiveness of the model are proved by using GAMS/BARON through the comparative analysis of four kinds of scenes. In order to further illustrate the practical application of the model, this paper applies the model to acrylic factory to optimize the alternative project of the factory and schedule the production schedule in the next month. The results show that considering the influence of value fluctuation and flexible time horizon will be more in line with the long-term development strategy of enterprises and bring greater benefits. The model also provides a new idea and method for dealing with the problem of project portfolio under uncertain conditions.

Key words: flexible time horizon, value fluctuation, project portfolio selection, binary tree