The herbicide imazethapyr was previously recommended for controlling weeds in soybean fields. However, the effects of imazethapyr on soil microbial communities and their relationship with crop root growth in sloped soils remain unclear. In this study, a field experiment was conducted on a sloped field to explore the effects of imazethapyr on crop root growth, microbial communities, microbial co-occurrence networks, and the interactions between microbes and crop root growth. The field experiment included two factors: slope and imazethapyr. The slope factor included three different slope gradients: 5° (S1), 10° (S2), and 15° (S3). The imazethapyr factor included two treatments: with (I1) and without (I0) imazethapyr. Thus, six total combinations of slope and imazethapyr treatments were tested in this study: S1I1, S2I1, S3I1, S1I0, S2I0, and S3I0. The results show that, compared to the I0 treatments, the I1 treatments significantly increased the soybean root length, surface area, and volume by 11.7~26.5 m, 171.7~324.2 cm2, and 1.8~3.1 cm3, respectively, across all the slopes. The Proteobacteria, Actinobacteriota, and Bacteroidota bacterial phyla and Ascomycota and Basidiomycota fungal phyla were found to be the top phyla represented bacterial and fungal communities. These five phyla were scattered in co-occurrence networks of bacterial and fungal communities, suggesting these phyla play critical roles in enhancing the stability of co-occurrence networks. Compared to the I0 treatments, the I1 treatments increased nodes from Proteobacteria, Actinobacteriota, and Bacteroidota phyla by 6.4%, 9.1%, and 11.2%, respectively, in the bacterial co-occurrence network. Similarly, in the fungal co-occurrence network, the I1 treatments improved nodes from Ascomycota and Basidiomycota phyla by 1.8% and 5.8%, respectively. Compared to the I0 treatments, the I1 treatments increased positive relations by 8.3% and 3.2%, respectively, in the bacterial and fungal co-occurrence networks. Moreover, the I1 treatments increased the relative abundance of root-promoting biomarkers and suppressed root-limiting biomarkers. However, the application of imazethapyr reduced the diversity and richness of bacterial and fungal communities in general. Furthermore, the nodes and links of bacterial co-occurrence networks in the I0 treatments were 9.2% and 78.8% higher than these in the I1 treatments. Similarly, the I1 treatments also decreased 17.9% of fungal community links compared to the I0 treatments. Our data also show that compared to the I0 treatments, the I1 treatments decreased almost all gene families encoding nitrogen and carbon cycling pathways. In conclusion, the application of imazethapyr increased soybean root growth by increasing root-promoting biomarkers and improved the stability and cooperation of co-occurrence networks of bacterial and fungal communities. However, the application of imazethapyr had some negative impacts on microbial communities, such as reducing the diversity of bacterial and fungal communities and nitrogen and carbon cycling pathways.