For many years, researchers have used the decomposition of grain‐size distributions to acquire critical information on provenances, transport dynamics and depositional environments. This study presents a novel decomposition method, termed the universal decomposition model, for analysing grain‐size data. The universal decomposition model unifies single‐sample unmixing and end‐member modelling analysis approaches and overcomes their respective limitations. To evaluate the effectiveness of the universal decomposition model, an artificial dataset and borehole data from the west Weihe Basin were analysed. Results indicate that the universal decomposition model algorithm performs proficiently on both datasets. Correlation analysis was employed to compare the abilities of universal decomposition model, single‐sample unmixing and end‐member modelling analysis to extract minor signals, with universal decomposition model and single‐sample unmixing exhibiting greater proficiency. Furthermore, the universal decomposition model provides a broader perspective for contrasting single‐sample unmixing and end‐member modelling analysis. The study highlights the inadequacy of the statistical method for determining the optimal number of components and summarizes an empirical approach. Moreover, disregarding the potential diversity in component shapes of real‐world sediments has been demonstrated to be a sub‐optimal design. Finally, this article presents results of a new investigation into the geological significance of sediment grain sizes revealed by various analytical methods that suggest that the universal decomposition model has enormous potential in reconstructing paleoenvironment.