Fresh, unpasteurized carrot juice is a popular element of the everyday diet of many consumers, and as such the matter of the juice's microbial safety remains an important one. Imaging flow cytometry (FCM) allows a fast enumeration and determination of cells, as well as their further differentiation. However, carrot juice is a difficult food product to analyze with the use of FCM due to interference from autofluorescence and the presence of plant debris. In this research, we aimed to obtain an effective and repeatable protocol for the preparation of carrot juice samples for FCM analysis. Through experimental and software-based means we successfully determined a reliable protocol for the preparation of fresh, unpasteurized carrot juice, which consisted of a sequence of filtering, centrifugation, enzyme treatment, and finally the implementation of the Machine Learning protocol for the best result.