Audio-assisted reading (reading-while-listening) was commonly used as a pedagogical method in English (L2) learning. Numerous studies had reported its efficacy in English (L2) reading. Its efficacy in reading comprehension has been inconclusive due to the lack of studies on the relationship among attention, cognitive load and L2 reading comprehension, with the possibility that the synchronous auditory input lessens attention to the visual input. We present a study of 41 Mandarin-speaking 8-year-old children reading English texts in three modes in a between-participants design. Data of cognitive load, comprehension scores and attention were fitted to a formal mathematical model, which confirmed that influences on L2 reading comprehension could be captured by interactions between attention and cognitive load. Based on the findings, three implications regarding how to appropriately apply auditory-assistant tools to L2 reading were generated.