Diagnosis of periodontal disease by Fourier transform infrared (FT-IR) microscopic technique was achieved for saliva samples. Twenty-two saliva samples, collected from 10 patients with periodontal disease and 12 normal volunteers, were pre-processed and analyzed by FT-IR microscopy. We found that the periodontal samples showed a larger raw IR spectrum than the control samples. In addition, the shape of the second derivative spectrum was clearly different between the periodontal and control samples. Furthermore, the amount of saliva content and the mixture ratio were different between the two samples. Partial least squares discriminant analysis was used for the discrimination of periodontal samples based on the second derivative spectrum. The leave-one-out cross-validation discrimination accuracy was 94.3%. Thus, these results show that periodontal disease may be diagnosed by analyzing saliva samples with FT-IR microscopy.