The purpose of this study was to detect Candida spp. on the tongue and in the subgingival sites in healthy and type 2 diabetes (T2D) patients with chronic periodontitis (CP), and to compare the accuracy of sampling methods. This study included 131 patients divided into four groups: healthy control (group A), nondiabetics + CP (Group B), diabetics with good metabolic control + CP (group C) and diabetics with poor glycoregulation + CP (Group D). Cotton swab samples from tongue and subgingival samples were obtained from each patient with help of sterile paper points and a sterile curette. Swab cultures were made on Sabouraud dextrose agar. The number of CFUs was counted. The sampling methods for subgingival plaque were compared by Receiving Operator Curve (ROC). The presence of Candida spp. on the tongue was statistically significant among groups (group D vs. others three groups: χ(2): p < 0.005 for each group). Positive findings of subgingival Candida spp. did not differ among the groups. There were no significant differences in the quantification of Candida spp., neither on the tongue, nor in the subgingival samples. 17.2% of diabetic patients revealed the presence of Candida spp. in the subgingival samples, with negative finding on tongue. There was a significant difference in the sampling methods for subgingival plaque (p = 0.000). Candida spp. is more prevalent on the tongue of diabetics. The sampling of subgingival plaque by a sterile curette is more accurate than with paper points. Subgingival plaque may represent a reservoir of commensals. It is necessary to standardize the sampling of subgingival plaque.