The traditional project management in the construction industry has the uniqueness of exclusive customization, the irreversibility of the project, the immobility of the project address, the long-term nature of the project, and the dynamic nature of the project. Based on these characteristics, the complexity and difficulty of project management in the construction industry are determined. Modern engineering is no longer satisfied with the traditional basic needs, but develops in a green, environmental friendly and efficient way. On the one hand, the project is required to meet good use requirements and meet the most basic functional requirements. On the other hand, the project construction process is required to meet the requirements of green environmental protection and ecological harmony. Therefore, the maintenance cost and operation cost corresponding to modern projects have become the key to the sustainable and healthy development of the engineering construction industry. Based on this, this paper will take the construction industry as an example to fully analyze the current situation and existing problems faced by the current project cost management. Through the research on the full life-cycle project cost management mode of the construction industry, the random matrix weight algorithm is creatively introduced, and the adaptive full life-cycle project cost management extended target tracking algorithm is proposed, which cuts the corresponding full life-cycle project cost management model into multiple submodels, and the corresponding management subobjectives are constructed from the submodel, and the expansion status of the corresponding subobjectives is described and analyzed in detail using the inverse distribution description, so as to accurately estimate the effect and target weight coefficient required by the management subobjectives, and finally realize the efficient, reasonable and scientific operation of the construction industry’s whole life-cycle project cost management, so as to maximize economic benefits. An experimental verification is carried out for a specific construction company. The verification results show that the full life-cycle project cost management scheme based on random matrix weight algorithm proposed in this paper has obvious advantages in management efficiency, cost control, process control, and other aspects compared with the traditional management scheme.