Banker plant systems increase biological pest control by supporting populations of natural enemies, i.e., using non-pest arthropod species as alternative prey. However, the presence of alternative prey does not always result in improved control of the target pest species owing to the complexity of biotic interactions. To increase the effectiveness of banker plants in IPM programs, a fine understanding of the indirect interactions between target aphid and alternative prey mediated by biocontrol agents is necessary. In this study, we first established a banker plant system, banker plant (Vicia faba)–alternative prey (Megoura japonica)–predator (Harmonia axyridis), to control the target pest (Myzus persicae) on pepper. We found that M. japonica strongly preferred faba bean as a host plant and posed no risk to Solanaceous crops. Harmonia axyridis adults had no significant predation preference for the alternative prey. In the short term, the interaction direction of the two aphid species depended on the relative initial density and the timescale. Harmonia axyridis showed a stronger negative effect on M. persicae than that on M. japonica. In the long term, the presence of alternative prey, M. japonica, enhanced the control effect of H. axyridis to M. persicae with initial density of 100–500 aphids per plant. The presence of the alternative prey could proliferate the population of H. axyridis, with from 0.2- to 2.1-fold increase of H. axyridis eggs. Overall, we put forward a strategy for setting the initial density of alternative prey of the banker plant system to target the high and low density of aphids, which highlighted the importance of indirect interactions in designing a proper banker plant system.