The care and interest for sustainable smart construction whose system consists of participants in construction, products and the process of construction, maintenance and management and their interrelationships is growing more and more stronger. Such a trend is the main driver for increased use of life cycle costing methods. While sustainable buildings have progressed in many aspects, it is financial sustainability that determines the extent to which a building is allowed to be sustainable. The problem of marina management stems partly from the lack of verified data needed for decision-making on maintenance and operation costs and future investments, their collection, processing and storage in databases, and partly due to different goals and conflicting general and individual interests of participants in the management process. What is also significant is the lack of a practical calculation of life cycle costs basis associated with determining the cost structure, as well as methods and tools for estimating these costs. Knowing the structure of marina life cycle costs is an important part of the marina management process, which should be based on scientifically researched and practically validated methods, especially in the beginning stages of conceptualization and selection of construction methods. A need to define marina life cycle costs that would be formulated through a general generic model based on the concept of life cycle cost analysis is observed during the research. The scope of the life cycle cost estimation model in this paper is limited to pontoons and the anchor system of pontoons and vessels. The primary goal of using the model is to estimate and optimize life cycle costs, so that different construction options can be quantitatively compared within the same decision-making process. The paper uses machine learning methods in order to examine the interrelationships of the collected data. The research includes determining all direct and indirect costs od construction, operation, maintenance, and end-of-life costs, designing basic machine learning models, comparing them and selecting the most favorable model with optimal selection of the number of input variables and optimization of model parameters depending on computer and data available resources. The obtained model will enable an increase in the level of decision-making quality and will provide a basis for model development of in the direction of quantitative comparison for different construction options.