In recent years, agricultural food safety incidents caused by poisonous pesticides have increasingly occurred in China, which has led to high levels of concern over the safety of agricultural products by the Chinese public. Thus, to solve the agricultural pest problem, there has been a movement toward the implementation of environmentally friendly agriculture, which benefits the health of both human and ecological communities. A number of studies have suggested that predators are the key natural factor for regulating and controlling pest populations. To help construct beneficial habitats in which predators can live and reproduce, scholars in countries outside China have conducted a large number of studies and trials on habitat restoration that focused on various issues, such as habitat selection and the diversity of predator species. However, few similar studies have been conducted in China, with few studies reporting how wind break forests regulate the ecological communities in agroforestry environments. This study evaluated the habitat selection and natural enemy function of the dominant predator in the agricultural fields in China, and analyzed the regulatory effects of shelter forests on the distribution of predatory insect populations in agroforestry environments. The aims of this study are to demonstrate that shelter forests could provide refuges and alternative food for the development and survival of natural enemy population, and to provide a scientific basis for the further investigation of natural enemy resources and proper utilization of non-crop habitats in agroforestry environments. This study investigated the habitat selection and natural enemy function of Chrysopa pallens Rambur, the dominant predator in the agricultural areas of Northern China. The investigation of habitat selection, oviposition sites, and overwintering sites in agroforestry systems including crop fields and shelter forests led to several conclusions. First, the adults of C. pallens were proficient in chasing aphid prey, and the larvae also exhibited clear spatial predation, capturing aphids at different spatial levels (spatial niche overlap index: P_{Larva·Spatial} > 0.65). The predatory stage of C. pallens was synchronized with aphid emergences (temporal niche overlap index: P_{Egg·Temporal} < 0.65). Second, forest belts adjacent to agronomic fields were the main overwintering places for C. pallens after the autumn harvest (P_α= 0.001, α = 0.05), which was important for maintaining the C. pallens population. Third, by cocooning as prepupae to overwinter in forest belts, the larvae of C. pallens primarily used bark cracks, topsoil, and fallen leaves and branches as overwintering places. There was no noticeable preference among these three types of overwintering habitats (P_α= 0.466, α = 0.05). Fourth, C. pallens migrated seasonally due to changes in habitat preference. During crop growth in summer and autumn, C. pallens migrated from forest belts to agronomic fields, preferring soybean and peanut habitats, where high-densities of aphids first appeared (LSD: P_α= 0.005, α = 0.01). After the fall harvest, and during winter and spring, C. pallens generally occupied forest belts, which were the only places left to overwinter after the crops were harvested. C. pallens preferred poplar coppice shoots, possibly because of the high aphid densities that occurred during early crop growth, which might serve as alternative food for survival and reproduction (LSD: P_α= 0.002, α = 0.01). Fifth, the preferred oviposition sites of C. pallens differed with habitat. In agronomic fields, eggs were laid from June to September, preferentially on peanut and corn plants, where high aphid densities would provide stable food sources for larvae (LSD: P_α= 0.007, α = 0.01). In forest belts, oviposition mainly occurred from May to July, preferentially in the poplar mixed forest belt, which provided a safe environment, in addition to be stable and diverse food sources (LSD: P_α= 0.038, α = 0.05). This study showed that both adults and larvae of C. pallens exhibit clear patterns of aphid chasing behavior. In addition, shelter forest habitats that were stable and complex under natural conditions might provide an important refuge for C. pallens in agroforestry environments when the landscape changed sharply after crop harvest. Thus, this study also implies that non-crop agroforestry system habitats are essential for the ecological regulation and management of agricultural pests because they provide a refuge for and improve the survival of natural enemies. Therefore, emphasis on cultivating forest belts in agricultural fields may be vital for the protection and maintenance of natural enemy populations in agricultural systems.