This study investigates how national identification, defined as patriotism, shapes the typical income-based pattern of support for redistribution in developing countries, with a focus on Southeast Asia. Using data from the Asian Barometer Survey (ABS) for eight Southeast Asian countries, we construct a multidimensional patriotism score based on four components, and measure redistributive preferences through four types of welfare-related policies. Redistributive preferences are conceptualized as support for welfare regime-building through post-market redistribution, including reducing the income gap between rich and poor and expanding access to healthcare, housing, and basic needs. Methodologically, we combine econometric models and Machine Learning (ML) techniques to analyze how income classes and patriotic attachment structure redistributive preferences. First, econometric results show that patriotism increases support for redistribution, but its effect is not strong enough to overturn the positive relationship between income and redistributive preferences. Second, in contrast, the ML models (Random Forest) highlight the high predictive importance of patriotism relative to socioeconomic factors. Third, we underscore the importance of country-level analyses, as both the strength of income cleavage and the influence of patriotism vary substantially across national settings. While the econometric framework reaffirms income-driven rational-choice mechanisms as the main structural divide, the ML results indicate that features grounded in social identity theory are powerful predictors for distinguishing individuals in terms of overall predictive contribution. This study contributes to understanding the complex interplay between civic national identification and income-based cleavage in developing contexts by mobilizing complementary approaches.