Owing to its great energy efficiency, dew point evaporative cooling is an ideal solution for cooling of electronics, data centers and electric vehicles, where a large amount of sensible heat is generated. To promote the application of dew point evaporative coolers, a common research gap between theoretical and experimental studies is addressed, i.e., how fundamental understanding can be turned into practical applications? In this paper, a coupled scaling and regression analysis is proposed as the key approach to linking the physics-based model to fast data-driven optimization. Accordingly, a complete model framework is developed for the dew point evaporative cooler by establishing a core regression model with its governing dimensionless numbers. The model is integrated with a robust multi-objective optimization algorithm for real applications. Instant predictions of product air temperature and maximum pressure drop can be obtained from the regression model, while it still retains some physical insights into how the cooling performance is affected by the dominant factors. A few optimization studies are carried out to navigate the optimal design and control strategies of the dew point evaporative cooler under assorted ambient conditions. It is noted that the regression model can accurately predict the experimental data of two coolers within ± 5.0% maximum discrepancy, and subsequent optimization suggests improved cooler designs with 30%–60% enhancement in energy efficiency, compared to an existing cooler prototype.