The high degree of microbial diversity found in soils is attributed to the highly heterogeneous pore space and the dynamic aqueous microenvironments. Previous studies have shown that spatial and temporal variations in aqueous diffusion pathways play an important role in shaping microbial habitats and biological activity in unsaturated porous media. A new modeling framework was developed for the quantitative description of diffusion-dominated microbial interactions focusing on competitive growth of two microbial species inhabiting partially saturated rough surfaces. Surface heterogeneity was represented by patches with different porosities and water retention properties, yielding heterogeneous distribution of water contents that varies with changes in relative humidity or soil matric potential. Nutrient diffusion and microbial growth on the variably hydrated and heterogeneous surface was modeled using a hybrid method that combines a reaction diffusion method for nutrient field with individual based model for microbial growth and expansion. The model elucidated the effects of hydration dynamics and heterogeneity on nutrient fluxes and mobility affecting microbial population growth, expansion, and coexistence at the microscale. In contrast with single species dominance under wet conditions, results demonstrated prolonged coexistence of two competing species under drier conditions where nutrient diffusion and microbial movement were both limited. The uneven distribution of resources and diffusion pathways in heterogeneous surfaces highlighted the importance of position in the landscape for survival that may compensate for competitive disadvantages conferred by physiological traits. Increased motility was beneficial for expansion and survival. Temporal variations in hydration conditions resulted in fluctuations in microbial growth rate and population size. Population growth dynamics of the dominant species under wet-dry cycles were similar to growth under average value of diffusion coefficients for dry and wet conditions, respectively, suggesting that the time-averaged diffusion coefficient could serve as a useful indicator for estimation of microbial activities in a highly dynamic system such as that found in soils.