Lodging is an important agronomic trait that affects soybean seed yield. In this study, a recombinant inbred line (RIL) population derived from ‘Zhongdou 27’ × ‘Jiunong 20’ (including 112 lines) was used to identify quantitative trait loci (QTL) associated with lodging of soybean. A genetic map of 2050.27 cM was previously constructed using 4412 single nucleotide polymorphism (SNP) bins in this population. Three major QTL were identified in the single environment for 3 years, accounting for 12.38–16.5% of the phenotypic variation. Among these QTL, qldg-1 was stable for 3 years and qldg-2 was stable for 2 years. QTL by environment interactions (QEI) mapping was also used to detect QTL. A total of 14 QTL were detected, which could explain 2.62–11.28% of the phenotypic variation. The constructed residual heterozygous lines (RHL) were used for the verification of qldg-1 and qldg-2, and the results showed that these two QTL could significantly improve lodging resistance. In addition, genes in the confidence interval of qldg-1 and qldg-2 were designed to predict the candidates. The results of quantitative real-time PCR (qRT-PCR) verification of five genes revealed that two genes (Glyma.17G048100 and Glyma.09G239000) were expressed differentially during the dynamic stages between the parents, demonstrating that these two were the candidates associated with soybean lodging. The QTL and candidate genes related to soybean lodging identified in this study will be of great significance to the future soybean molecular-assisted breeding for lodging resistance.