P-s-n curves detailing fatigue are basic data required to ensure high reliability design. Since fatigue testing requires significant time and resource costs, the development of reliable small-sample methods is crucial to ensure the reduction in time and costs of design process. This paper investigates the possibility for applying B-spline interpolation basis functions to approximate the probability density functions on various levels of high-cycle uniaxial fatigue using the sample data of only selected stress levels. It was found that using the proposed model an iterative approach can be used to ensure acceptable fitting accuracy and to predict fatigue life on the desired probability level. As a result, the findings show the possibility for rapid modeling of the probability density function and the resulting reliability in high-cycle fatigue by using the well-known interpolation techniques based only on the basic dispersion parameters. The proposed approach provides a robust framework for fatigue life prediction, paving the way for broader applications in engineering analysis. The proposed data augmentation method seems to demonstrate the potential to reduce fatigue testing costs and simplify testing procedures.