This study analyzed the relationship between frequent policy transitions between redevelopment and regeneration projects and the physical characteristics of parcels in low-rise residential areas of Seoul. It focused on twelve neighborhood-type urban regeneration areas where redevelopment cancellation, regeneration designation, and recent redevelopment re-designation have overlapped. Parcels were categorized into four groups: Regeneration only, Redevelopment → Regeneration, Regeneration → Redevelopment, and Redevelopment → Regeneration → Redevelopment. Using parcel-level physical characteristics, this study conducted ANOVA and chi-square tests to examine group differences, followed by a random forest model to identify the relative importance of variables in distinguishing policy transition groups. The results indicated statistically significant differences in land area (parcel size), topography, building age, zoning, and parcel shape across the groups. Random forest analysis identified these factors as the most influential for classification, with higher accuracy in distinguishing Groups 1 and 4. However, Groups 2 and 3 exhibited weaker separability. These findings empirically support two hypotheses: first, parcel-level physical characteristics differ across policy transition types in areas experiencing frequent shifts between redevelopment and regeneration; and second, under the assumption of constant socio-economic conditions, physical characteristics serve as key determinants of policy type. These results highlight the need to consider parcel-scale physical attributes when formulating redevelopment and regeneration strategies, while acknowledging that additional socio-economic and institutional factors interact with physical conditions to shape policy transitions.