Information on the food habits of wild Japanese macaques (Macaca fuscata) is typically obtained from behavioral observation. Dietary information obtained from feces, for which few Japanese primatologists have paid attention, is also useful. Collecting information on dietary habits through both fecal analyses and behavioral observation enables us to understand the value of both methods in assessing their ability to obtain dietary information. If we apply this relational equation to unhabituated animals, such as crop feeding ones, we can estimate the feeding behavior of unhabituated animals through the fecal analyses. To test this idea, we analyzed fecal contents of well-habituated wild Japanese macaques from which we had also obtained behavioral data. We then preliminarily conducted statistical modeling for the consumption of target food categories from fecal contents. For most feeding parts, we found that presence of the target category in fecal samples positively affected % value of that food in the diet. Models obtained for bark and flowers were well-fitted, while models for other main items (leaves, berries, nuts, and herbs) were less well fitted. This is likely due to differences in the effectiveness of detecting different food contents with these two methods. In the case of nuts, their higher digestibility could affect detectability leading to a reduction in the models precision. Our study confirmed the effectiveness of fecal analyses as a tool for evaluating feeding behavior, but we need to consider other factors (such as digestibility and season) to improve the accuracy of this model.