The reliability and availability of nuclear power plant (NPP) functions is related to the operational safety of NPP, licensee requires a robust categorization process to carry out safety classification of functions. The safety classification of functions is typically conducted using qualitative or quantitative methods to evaluate their impact on accident consequence mitigation and reactor safety control. This study primarily analyzed two risk-informed methods for evaluating the safety significance of functions by probabilistic safety assessment (PSA), which respectively uses basic events and fault trees/gates to represent the function's operation. This paper conducts a pilot application based on the CNP1000 unit and includes comparative analysis of the methods. Based on the analysis results, it can be concluded that the risk-informed approach using PSA model to quantitatively calculate the function importance, and comparing it with importance measure criteria, enables a rational classification of NPP functions into high safety significance (HSS) and low safety significance (LSS). Additionally, the analysis results indicate that differences in the calculated importance of functions arise due to the use of varying representative objects for function across different methods. Finally, the applicability analysis of methods is conducted for different system functions and application scenarios, which analyze the modeling methods of function in PSA model and the characteristics of importance measure, and provide methods recommendations and conclusions to improve the accuracy of safety classification results.