Background/Objective: Cervical intraepithelial neoplasia grade 2 (CIN2) shows heterogeneous clinical behavior, with substantial rates of spontaneous regression under active surveillance. Reliable molecular biomarkers are needed to distinguish regressive from transforming lesions and reduce overtreatment. We evaluated the prognostic role of host and viral DNA methylation, alone and combined with HPV genotyping, in predicting CIN2 regression. Methods: This subanalysis derives from a prospective, multicenter Italian cohort of women with histologically confirmed CIN2 managed conservatively. Among 319 enrolled women, 134 with single HPV infections and valid host (FAM19A4/miR124-2) and viral (HPV L1 region) methylation results were included. HPV genotyping was performed with partial stratification (HPV16/18 vs. non-16/18). Clinical outcomes at 24 months were classified as regression versus persistence/progression. Logistic regression models assessed associations between biomarkers and regression. Results: At 24 months, 50% of women showed regression. Host and viral methylation positivity rates were more frequent in non-regressive lesions (40.3% vs. 19.4%, p = 0.01, and 52.2% vs. 32.8%, p = 0.02, respectively). Negative host methylation was significantly associated with regression (Odds Ratio OR = 0.37, 95% CI 0.17–0.81, p = 0.02), as was negative viral methylation (OR = 0.47, 95% CI 0.23–0.96, p = 0.04). Conclusions: Both host and viral methylation are inversely associated with CIN2 regression. Combining methylation markers did not substantially improve predictive accuracy; however, methylation negativity emerged as a potential molecular reassurance marker. When integrated with HPV genotyping, the highest probability of regression was observed among women with non-HPV16/18 infections and negative methylation results. These results endorse DNA methylation testing as a molecular tool for the conservative management of CIN2.