Changes in farmland landscape patterns can significantly affect the occurrence of and damage caused by pests. Different landscape patterns have different effects on pest populations, and thus elucidating the ecological effects of farmland landscape patterns on pests is an important factor in pest control. In this study, we chose the cotton planting district of Shandong Province in North China as our research area. We selected 14 sample counties or districts, which were as close to different land coverage types as possible. We determined population size during the primary period of cotton aphid damage. Aphids were on cotton seedlings in mid-to late May, whereas summer aphids peaked at approximately June 25. Landscape indices for each county or district were calculated from the distribution of land coverage types, which were obtained by comprehensively analyzing data from remote satellite sensing. Populations of cotton aphids were also surveyed in the corresponding counties or districts. We analyzed the correlation of multi-factors, including landscape composition, landscape quantity, and landscape configuration with cotton seedling aphids and summer aphids at the provincial level. The results showed that landscape patterns had a close association with both types of cotton aphids, which had significantly different population sizes in the same counties, whereas there were different response characteristics for cotton seedling aphids and summer aphids. Cotton seedling aphid populations has a significant positive association with total area, the fractal dimension index of farmland, contiguity index of the county, and the county's radius of rotation, and were negatively correlated with Simpson's diversity index. Summer aphid populations had a positive relationship with patch richness density and the contiguity index of residential industrial traffic. In general, responses of cotton seedling aphids and summer aphids to the contiguity index were similar, and there were more cotton seedling aphids and summer aphids in landscapes with less fragmentation. The responses of cotton seedling aphids to landscape diversity were different than those of summer aphids, and the diversity of the landscape benefited the summer aphid population, whereas it had a non-Significant relationship with seedling aphids. Patch richness density could significantly affect cotton seedling aphid populations, but not summer aphids. These results indicated the complexity of the responses of insect pests at different stages to landscape patterns. Therefore, in regions with a higher quantity of cotton aphids, we can control the population quantity of cotton aphids by reducing wintering sites, decreasing the contiguity index of residential industrial traffic, and increasing Simpson's diversity index according to the occurrence of cotton seedling aphids and summer aphids. Furthermore, we could form a regional comprehensive strategy for the control of cotton aphids by combining other biological and chemical control methods. As such, this is the first report, to our knowledge, that demonstrated that the same species occurring at different stages could have different responses to landscape patterns, which may be caused by different quality and quantity of food resources, different temperature and humidity, or other environmental conditions. We have provided theoretical and practical foundations to support the regional control of cotton aphids.