Land use, local management, and seasonal variation significantly impact the ecological dynamics of bee–plant communities and their ecological interactions. These variables negatively affect diversity and ecological interaction networks within human-dominated landscapes. Additionally, seasonal variables such as temperature, rainfall, and resource availability across different seasons play essential roles in shaping bee communities and their interactions with flowering plants. However, little is known about how diversity and ecological interaction networks of non-crop plants in agricultural landscapes respond to intra-seasonal variations, specifically within the rainy season. In this study, we assessed how land use types, coffee crop management, and intra-seasonal variation within the rainy season influenced the composition and diversity of bee and plant communities, and their interaction networks in semi-natural habitats surrounding coffee plantations. We recorded the diversity of bees and plants and analysed their interactions networks metrics, such as specialisation, nestedness, modularity, connectance and bee/plant generality, in 8 pairs of sites. Our findings indicate that human settlements negatively influence bee generality, suggesting that human-dominated land and the introduction of exotic plants reduce floral resources for bees, which may decrease bee visitation. In contrast, extensive semi-natural and forested areas seemed to support bee generality. Additionally, we observed higher visit frequency and richness of bees and plant generality during the second period of the rainy season (July to October), leading to more robust bee–plant interaction networks in the same period. This study enhances our understanding of how land-use types and intra-seasonal climatic variation shape structure of bee floral visitor communities and their interactions with flowering plants. Furthermore, our findings underline the negative impact of human-dominated landscapes on the ecological dynamics of plants visited by bees and their interaction networks.