Laser induced breakdown spectroscopy (LIBS) has become an important analytical tool for extracting cultural relic information in the archaeological field because of its rapid analysis and micro damage. In this study, random forest (RF) and LIBS method were used to quantitatively analyze the brightness and identify the source of white porcelain in Xi'an. Firstly, the brightness of white porcelain fragments is measured, and the spectral brightness of white porcelain samples was analyzed quantitatively by LIBS platform. D2nd-VIM-RF calibration model was established for multi-element quantification, achieving excellent prediction performance with an R2CV of 0.9814 and an RPD of 4.6. Secondly, an RF classification model was developed to distinguish two types of white porcelain from Xi'an. Based on spectral data, the model was optimized using MSC-WT preprocessing and VIM-SPA variable selection, resulting in the MSC-WT-VIM-SPA-RF classification model. It achieved a sensitivity of 0.9670, specificity of 0.9438, and overall accuracy of 0.9600, significantly enhancing classification reliability. To optimize both the predictive capability and explanatory power of the model, quantitative brightness was incorporated as an auxiliary variable alongside LIBS spectra. Brightness reflects surface features such as glaze quality and microstructure, which are closely associated with raw material composition and firing techniques critical factors for distinguishing porcelain types. This study underscores the significant potential of combining LIBS and RF technology for rapid, minimally invasive analysis of ancient white porcelain, offering valuable insights into the classification, identification, and quantitative analysis of historical ceramics.