Habitat Suitability Index (HSI) has been modeled to predict the habitats of wild animals. In this research we focus on red squirrelSciurus vulgaris orientis living in Hokkaido. To identify mammals to visit the research sites, the near infrared red cameras have been installed randomly over Nakagawa experimental forest of Hokkaido University since 2008. It is expected to get the relation between the forest structure and the spatial usage of squirrels. To achieve this goal, a poisson regression is used to analyze the relationship. To make a model, the number of squirrel’s visits is used as a response variable, and the tree parameters are used as predictor variables. The model that has the lowest value of Akaike’s Information Criterion (AIC) is accepted as the best model. As a result, the variables selected by this process to make the final model were stem density of coniferous trees, the number of tree species, and coefficient variance of tree height. From this result, red squirrels tend to live in coniferous dominant forest, the sites have the various tree species as food resource all the year round, and various height of trees. This result enables us to understand that red squirrels like the forest structure. The potential habitat area from this model can be used to prevent the road crossing accidents.