This paper investigates the application of fuzzy sets in the classification of metagenomic data, particularly focusing on the microbiomes of animal intestines in various urban environments. The study aims to assess how the utilization of fuzzy techniques impacts classification efficiency. The dataset comprises sequencing data obtained from rodents inhabiting different urban green areas and control agricultural regions. Fuzzy modification of the three-sigma rule is employed to express numerical variables as fuzzy sets, allowing for the introduction of descriptors describing each element’s relationship to others within the variable. Experimental results demonstrate that the selection of elements for the training set significantly influences classification efficiency. The study compares classification based on raw data with that based on fuzzy descriptors and a combination of both. Results indicate that while fuzzy descriptors enhance classification efficiency, excessive fuzzification does not necessarily improve results. The optimal approach involves combining raw data with a moderate level of fuzzification. The study concludes by discussing future research directions, including the comparison of different classification methods and aggregation techniques.