Abstract Ecological intensification (EI) provides an important and increasingly adopted pathway for achieving more sustainable agricultural systems. However, the implementation and success of on‐farm EI practices may vary depending on landscape context and local management practices. We evaluated how EI interventions, including two different agricultural input regimes (high or low use of synthetic pesticides and fertilizers) and floral border crops, affected local natural enemy biodiversity, pest abundance and crop yield, and how this was influenced by the surrounding landscape context across 12 rice fields on smallholder farms in Puducherry, India. Reducing agricultural inputs positively impacted the overall natural enemy assemblage; however, responses to landscape factors varied. For example, coccinellid beetles were negatively correlated with higher densities of field edges (landscape configuration). In contrast, spiders, the most abundant group surveyed, were not significantly influenced by any landscape metric. Furthermore, pest abundance was greatest in fields with reduced inputs but only at sites where floral border crops were not present. Mean rice grain yield was lower across low‐input sites compared with high‐input sites and floral border crops had opposing effects across high‐ and low‐input sites. At low‐input sites, mean yields were 33% higher where floral border crops were present. At high‐input sites, the presence of floral border crops was correlated with a lower mean yield (16%). These findings show that ecological intensification practices can benefit smallholder crop systems but highlight the need to account for variations in landscape context and local management practices for developing effective sustainable management practices.