Species co-occurrence patterns are widely used to infer the ecological suitability of species that are absent from local communities. However, such approaches are often framed within the concept of dark diversity—defined as the set of species that are ecologically suitable but currently absent from a local community—their predictions reflect different underlying mechanisms related to species co-occurrence and regional frequency. In this study, we compared two co-occurrence-based methods, Beals' index and the hypergeometric method, using vegetation survey data from mixed broadleaved-Korean pine forests (MBKF) across different successional stages in Northeast China. Method performance was assessed by predicting species suitability from co-occurrence patterns and validating predictions against observations from the surrounding area. The results show that both methods effectively assign ordered suitability values consistent with species occurrence status. Beals’ index exhibited higher overall predictive accuracy but showed greater variability among plots. In contrast, the hypergeometric method provided more stable performance and yielded suitability estimates that were ecologically informative for rare species. These findings demonstrate that co-occurrence-based suitability estimates are highly sensitive to the assumptions inherent in each method. Consequently, method selection should be guided by specific research questions and management objectives. Such careful methodological choice is crucial for deriving reliable conclusions and for effectively applying co-occurrence-based approaches in biodiversity assessment and forest management.