Sampling site maps not only show intuitively distribution of the sampling sites where soil samples have been collected, but also provide reference for researchers to determine where supplementary soil samples should be taken. Traditionally, in soil mapping, codes were used to label each sampling point in the map for identification. However, as affected by their spatial geographic locations, it is impossible to label the sampling sites in sequence, thus making it very hard to locate a required sampling site in the map, especially when it has too many sampling sites located densely or is too large in format. To realize orderly labeling of sampling points, a “Soil Sampling Point Labeling Model for thematic soil maps (SAMPLA)" was developed, consisting of three sub-models, i.e. map reading vision zone division model, soil sampling site attribution map reading vision zone determination model and, sampling point sequential labeling model. With the aid of ArcGIS 10.0 and Human Machine Interaction（HMI）design, SAMPLA was translated into a computer program with the C# programming language, and tested with soil sampling points in the 1:50 000 national standard map sheets and county-level maps. Results show that the model is applicable to maps different in type and scale, helps realize standardized mass mapping, and improves map reading efficiency. But labeling was slightly affected by the map reading vision zone division program. However, besides in soil sciences, SAMPLA can also be used in other scientific fields, such as environmental science, water science, and geological science, for making sampling point distribution maps.