Long-term application of chemical fertilizers causes soil degradation and nitrogen (N) loss, but these effects could be alleviated by organic fertilizers. In addition, crop rotation is a feasible practice to increase soil fertility, soil quality and crop yields comparing with monocultural cropping patterns. However, questions remain concerning how the soil microbiome responds to different manure application rates under crop rotations. Here, we collected soil samples from a rice-rape system to investigate the response of the soil microbiome to nine years of pig manure application at different rates (CK: 0 kg ha^-1, M1: 1930 kg ha^-1, M2: 3860 kg ha^-1 and M3: 5790 kg ha^-1). Our results revealed that the bacterial α-diversity (Chao1 and Shannon index) in the rape season increased first and then decreased with increasing manure application rates, and a high manure load tended to decrease the bacterial α-diversity in the rice season. Long-term manure application enriched some copiotrophic bacteria, such as Proteobacteria and Actinobacteria, while it decreased the relative abundance of Nitrospirae. Redundancy analysis (RDA) and the Mantel test indicated that soil pH, TC, TN, AP, C/P and N/P ratios were the main factors influencing bacterial communities. Moreover, network analysis showed that a low manure application rate shaped a complexly connected and stable bacterial community, while higher manure application rate decreased the stability of the bacterial network. These findings improve our understanding of bacterial responses to long-term manure application under crop rotations and their relationships with soil factors, especially in the context of increasing fertilizer inputs.