Aim To evaluate the cross‐sectional association between severe periodontitis and diabetes mellitus (DM), in a representative sample of Spanish population. Materials and Methods The di@bet.es epidemiological study is a population‐based cohort study aimed to determine the prevalence and incidence of DM in the adult population of Spain. The at‐risk sample at the final examination (2016–2017) included 1751 subjects who completed an oral health questionnaire. This questionnaire, together with demographic and risk factors, had been previously validated to build an algorithm to predict severe periodontitis in the Spanish population. Logistic regression models were used to evaluate the association between severe periodontitis and DM with adjustment for confounding factors. Results In total, 144 subjects developed DM, which yielded 8.2% cumulative incidence. Severe periodontitis was detected in 59.0%, 54.7% or 68.8% of the subjects depending on three different selected criteria at the 2016–2017 exam. All criteria used to define severe periodontitis were associated with DM in unadjusted analysis, but the magnitude of the association decreased after adjusting for significant confounders. The criteria ‘≥50% of teeth with clinical attachment loss ≥5 mm’ presented an odds ratio of 4.9 (95% confidence interval: 2.2–10.7; p ≤ .001) for DM. Conclusions Severe periodontitis is associated with DM in the Spanish population.