Purpose. Increasing the level of diagnostics of the technical condition of nuclear power plant power units by improving the methodology for continuous monitoring and forecasting the residual resource of technological equipment using digital twin models built on the basis of fractal-cluster analysis methods. Methodology. The following methods were used in the research process: fractal-cluster (for clustering data from sensors monitoring technological parameters); system-structural approach (for building a control system: power plant, power unit, technological process parameters); simulation modeling (for modeling digital twins based on graph structures and hierarchical data models). Findings. An improved method of continuous monitoring, control, diagnostics, assessment and forecasting of the state of the residual resource of technological equipment of power units of nuclear power plants using digital twin models is proposed. The value of the integrated index of quantitative assessment of the technical condition and the time of achieving the loss of functional suitability of the technological equipment of the power unit of the nuclear power plant are taken into account. Originality. Unlike traditional methods for assessing the condition of a nuclear power plant unit, the proposed method assesses the condition of the technological equipment of the power units using digital twin models. It allows taking into account the multilevel and nonlinear nature of the degradation process and the reduction of the residual resource of the equipment. A system of criteria and indicators of the integrated technical condition index based on fractal and time characteristics has been formed. This allows for prompt (in real time) response to early manifestations of defects and the increased accuracy of long-term forecasting of the resource of the technological equipment of the nuclear power plant unit. Practical value. The proposed method of continuous monitoring of the state of technological equipment of nuclear power plant power units using digital twin models ensures timely detection of defects. It increases reliability and safety, optimizes maintenance schedules and reduces the time for diagnosing emergency risks. This approach can become the basis for creating intelligent systems for continuous monitoring and life cycle management of power equipment.