In this paper, for the information categorization of toponyms in ethnic minority agglomerations, different cultural partitioning results are derived by combining the region of interest model, which constitutes a specific semantic map of the region of interest model, i.e., a semantic map of the toponymic structure. The preference index of toponymic culture is taken as a feature variable, and the calculation of toponymic similarity in ethnic minority agglomerations is realized by adopting the method based on the difference of feature vectors. Dynamically determining the weights is used to evaluate the semantic similarity of spatial data for place names. To examine the distribution characteristics and distribution patterns of place names in ethnic minority agglomerations, kernel density and logistic regression models are utilized. The ratio of the number of place names of each ethnic group to the total number of county units is analyzed by Moran’s I index to obtain the spatial autocorrelation distribution direction of the distribution of place names in Yunnan Hani Autonomous County. The analysis shows that the spatial autocorrelation index of the Hani ethnic group is 0.3524, which mainly falls in the first, second and third quadrants, where the first quadrant indicates that the area around the region with a higher proportion of Hani ethnic group place names is likewise a positively correlated region with a higher proportion of Hani ethnic group place names. As per the semantic map analysis results, the spatial distribution of the population closely corresponds with the distribution of toponyms in ethnic minority areas.