Food web structure and the trophic interactions nested within, serve the basis of many ecosystem functions in ecological communities. In agriculture, the balance of these functions (e.g. biocontrol, primary production, etc.) is essential for optimizing production. However, studies on food web structure have primarily relied on co-occurrence data for inference which limits research to specialist predators (i.e. parasitoids) that have more predictable species interactions. To fill this gap, we employ a stable isotope approach to examine how food web structure of generalist predators are shaped in paddy fields. In addition, we explored the possibility that food web structure may be influenced by farming method and landscape. Our results show that as the crop develops, predators significantly shifted toward a rice-based diet, indicated by changes in their carbon isotope signatures. Moreover, arthropod communities also accumulated more species over time and this increase predicted indices of community trophic diversity (number of trophic levels, basal resource utilized, and niche width). However, we did not find any statistically significant differences in the food web structure of generalist predators between farms operating under different farming methods and landscape contexts. As we increase our sample size in the future by adding more farms and collecting long-term data, we hope to increase statistical power and the robustness of our study. Nevertheless, the results of our study offer a relatively simple method to decipher the complexity of (semi-)natural food webs using stable isotopes.