Many passerine bird populations, particularly those that have open-cup nests, are in decline in agricultural landscapes. Current theory suggests that an increase in habitat generalist predators in response to landscape change is partially responsible for these declines. However, empirical tests have failed to reach a consensus on how and through what mechanisms landscape change affects nest predation. We tested one hypothesis, the Additive Predation Model, with an artificial nest experiment in fragmented landscapes in southern Queensland, Australia. We employed structural equation modelling of the influence of the relative density of woodland and habitat generalist predators and landscape features at the nest, site, patch and landscape scales on the probability of nest predation. We found little support for the Additive Predation Model, with no significant influence of the density of woodland predators on the probability of nest predation, although landscape features at different spatial scales were important. Within woodlands fragmented by agriculture in eastern Australia, the presence of noisy miner colonies appears to influence ecological processes important for nest predation such that the Additive Predation Model does not hold. In the absence of colonies of the aggressive native bird, the noisy miner, the influence of woodland predators on the risk of artificial nest predation was low compared with that of habitat generalist predators. Outside noisy miner colonies, we found significant edge effects with greater predation rates for artificial nests within woodland patches located closer to the agricultural matrix. Furthermore, the density of habitat generalist predators increased with the extent of irrigated land-use, suggesting that in the absence of noisy miner colonies, nest predation increases with land-use intensity at the landscape scale.