Human serum albumin (HSA) injectable product is a severely afflicted area on drug safety due to its high price and restricted supply. Raman spectroscopy performances high specificity on HSA detection and it is even possible to determine HSA injectable products noninvasively. In this study, we developed a noninvasive rapid screening method for of HSA injectable products by using portable Raman spectrometer. Qualitative models were established by using principal component analysis combined with classical least squares (PCA-CLS) algorithm, while quanti-tative model was established by using partial least squares (PLS) algorithm. Model transfer in different instruments of both the same and different apparatus modules was further discussed in this paper. A total of 34 HSA injectable samples collected from markets were used for verification. The identification results showed 100% accuracy and the predicted concentrations of those identified as true HSA were consistent with their labeled concentrations. The quantitative results also indicated that model transfer was excellent in the same apparatus modules of Raman spectrometer at all concentration levels, and still good enough in the different apparatus modules although the relative standard deviation (RSD) value showed a little increasing trend at low HSA concentration level. In conclusion, the method was proved to be feasible and efficient for screening HSA injections, especially on its screening speed and the consideration of glass containers. Moreover, with inspiring results on the model transfer, the method could be used as a universal screening mean to different Raman instruments.