To enable rapid prediction of the skin-whitening efficacy of dandelion ethanol extract, quantitative prediction models for ABTS and DPPH radical scavenging activities, tyrosinase inhibition rate, and whitening activity index were developed using ultraviolet-visible (UV-Vis) spectroscopy combined with chemometric methods. The underlying mechanisms were further investigated through Fourier transform infrared spectroscopy (FT-IR), liquid chromatography-tandem mass spectrometry (LC-MS/MS), and molecular docking studies. The results demonstrated that UV-Vis spectroscopy combined with feature extraction exhibited robust predictive performance for all whitening-related indicators. The optimal prediction correlation coefficients (Rp2) for each indicator exceeded 0.85, and the residual predictive deviation (RPD) values in the validation set were greater than 2.5. Compared to raw spectra, the application of feature extraction could significantly enhance the performance of predictive models. The functional group information obtained from FT-IR, characteristic ultraviolet absorption wavelengths, and LC-MS/MS analysis collectively indicated that flavonoids and phenolic acids were the primary active components contributing to the whitening efficacy. Molecular docking analysis revealed that flavonoids and phenolic acids exhibited stable binding interactions with key target proteins, indicating their potential role in inhibiting melanin synthesis and contributing to the whitening efficacy. This study presented a rapid and reliable approach for evaluating the whitening activity of dandelion ethanol extract. Furthermore, the underlying mechanisms were elucidated through an integrated analysis combining spectroscopic characterization, component identification, and molecular mechanisms. These findings could provide methodological support for the efficient evaluation of whitening efficacy in traditional Chinese medicine and contribute to a deeper understanding of its mechanistic basis.