<p>Banker plants, non-crop plants that provide supplemental resources for natural enemies, have been used to augment the efficiency of biological pest control. While many studies have individually examined parameter values of pests and natural enemies (e.g., demography, predator–prey interactions), how such factors jointly determine the ultimate success or failure of biological control has not sufficiently been examined due to its experimental difficulties. In this study, we constructed a population dynamics model comprising a single pest and natural enemy species with two distinct patches (crop and banker plants). We assumed that only natural enemies could reciprocally move between patches and that the substantial population size of natural enemies entering each patch varied depending on the banker-to-crop area ratio. Considering the whitefly <i>Bemisia tabaci</i> as the pest and <i>Nesidiocoris tenuis</i> as the natural enemy, we applied the parameter values obtained in previous empirical studies. Numerical simulations revealed that using the banker plant immediately made the pest density low and stable, whereas damped oscillations of pest density were observed in the absence of bankers. We also showed that the larger the outflow of the natural enemy, the higher the pest density, which eventually became higher than that without a banker or economic injury level. However, even under such conditions, a larger banker area can compensate for the negative impact of outflow and augment the efficiency of pest control. However, we also showed that too high performance of the banker can lead to non-target effects by zoophytophagous natural enemies, such as <i>N. tenuis</i>.</p>