Banker plant systems have the advantages of introducing natural enemies preventively and maintaining by providing alternative prey, thus controlling the pests sustainably. Banker plant systems are usually composed of three factors: a banker (secondary) plant, an alternative prey, and a shared predator (attacking an alternative prey on the secondary plant, and the targeted pest on the crop). However, for most banker plant systems, there is a lack of understanding regarding the dynamic relationship among these elements, with detrimental effects on practical applications. Therefore, in this study, the control of Myzus persicae on Capsicum annum by the Coccinella septempunctata–Megoura japonica–Vicia faba banker plant system was used as the research system. The effects of different release time of predators, different initial numbers of alternative prey (Me. japonica) and different initial ratios of target pests/alternative prey on the indirect interaction of two aphids and the biological control effect of shared predators were tested. The occurrence of indirect interactions between the two aphid preys, the impact on population dynamics, and biological control effectiveness of the shared predator C. septempunctata were investigated. When the initial numbers of both species of aphids were equal (200 each), the delay between aphid and C. septempunctata introduction in the cage had no effect on My. persicae, but Me. japonica showed lower numbers when testing the least time between predatory introductions. When the numbers of the two aphids were manipulated, the My. persicae population was significantly reduced by the predator only at a ratio of My. persicae to Me. japonica < 1, while initial ratios ≥ 1 enabled My. persicae population growth. In 1–6 days, the control effect of C. septempunctata was the best. Principal component analysis showed that the experimental time, initial numbers of Me. japonica, and relative numbers of Me. japonica affected the predation of My. persicae by C. septempunctata. In addition, when the initial aphid ratio was greater than 1/4, C. septempunctata was able to effectively control My. persicae. Overall, our study confirmed the number-mediated indirect interaction (apparent competition) relationship and its impact on prey population dynamics. We provide useful information for optimizing banker plant systems, to boost biocontrol of aphid pests.