**Background:** Continuous soybean monoculture in northeastern China has led to soil acidification, increased soil-borne diseases, decreased enzyme activity, and toxin accumulation. Crop rotation is a well-known management practice to mitigate these issues, but the effects of specific cropping sequences on soil microbial communities and their relationship to soil properties and crop yields remain poorly understood, particularly under conventional fertilization. This study aimed to determine whether planting sequence affects soil microorganisms and whether core beneficial microorganisms can increase crop yields.
**Methods:** A long-term field experiment was established in spring 2012 in Heihe City, China (E 49°33′35″, N 125°27′5″; 225 m a.s.l.; dark brown soil) under a mid-temperate semi-humid continental monsoon climate (mean annual temperature 0.5°C, annual precipitation 481 mm). Six planting sequences were tested over five years: continuous soybean (sss), wheat-soybean (wsw), wheat-maize-soybean (wms), soybean-maize (smm), maize-soybean (mss), and continuous maize (mmm). Each treatment had three replicate plots (87.75 m² each). Fertilizer was applied as base fertilizer at rates of N-P₂O₅-K₂O 135-67.5-45 kg/ha for maize, 50-60-45 for soybean, and 75-80-45 for wheat. After five years, crops were harvested and yields calculated by dry weight. Soil samples (0–15 cm depth, five subsamples per plot mixed into one bulked sample) were collected. Soil properties (pH, moisture, TOC, TN, TP, TK, AN, AP, AK) were measured. DNA was extracted using the PowerSoil DNA isolation kit. Bacterial 16S rRNA genes (V3–V4 region, primers 341F/806R) and fungal ITS region (primers ITS1F/ITS2R) were amplified and sequenced on the Illumina HiSeq 2500 platform. Sequence data were processed using Trimmomatic, FLASH, UCHIME, and UPARSE (97% OTU clustering). Taxonomy was assigned using RDP (bacteria) and UNITE (fungi) databases. Alpha diversity (Chao, ACE, Shannon, Simpson) was calculated in Mothur. PCoA, RDA, network analysis, PICRUSt functional prediction, and FUNGuild analysis were performed in R.
**Key Results:** A total of 799,070 bacterial and 790,459 fungal high-quality reads were obtained, clustered into 7,564 bacterial OTUs and 2,895 fungal OTUs. All crop rotations increased soil TOC, TN, and AN compared to monocultures. Soybean-maize rotation (mss) increased soybean yield by 11.27% compared to continuous soybean (sss). Compared to continuous maize (mmm), wms and smm increased maize yield by 1.64% and 23.06%, respectively, with smm being the most significant. The wms system maintained near-neutral soil pH, while wsw resulted in soil pH similar to soybean monoculture. The number of observed bacterial species was highest in smm and lowest in mss; conversely, fungal species were highest in mss. Bacterial Shannon index was significantly positively correlated with soil TK (p < 0.05), and Simpson index was significantly positively correlated with TN (p < 0.05) and AP (p < 0.01). Fungal Shannon and Simpson indices were significantly positively and negatively correlated with AK (p < 0.05), respectively. RDA showed that TOC, TN, and TP significantly affected bacterial community structure, while soil pH was the main factor affecting fungal community structure. The dominant bacterial phyla across all treatments were Proteobacteria, Acidobacteria, Actinobacteria, Chloroflexi, and Gemmatimonadetes. The relative abundance of Proteobacteria was higher in mss and smm. The sss system had high Acidobacteria abundance. At the fungal level, Ascomycota and Basidiomycota dominated. FUNGuild analysis revealed that the proportion of animal and plant pathogens followed the order: mss < mmm < smm < wsm < wsw < sss. Beneficial microorganisms (endophytes, ectomycorrhizal, arbuscular mycorrhizal fungi) were highest in mss, followed by smm, and lowest in wms. Network analysis showed Acidobacteria and Actinobacteria had positive correlations with other members, while Proteobacteria mostly showed negative correlations.
**Clinical Implications:** This study provides evidence that soybean-maize rotation sequences (particularly smm and mss) are superior management strategies for improving soil fertility, increasing crop yields, and promoting a healthier soil microbiome in the major grain-producing regions of northeast China. The smm system produced the highest maize yield (23.06% increase over continuous maize), reduced pathogen abundance, and increased beneficial microbial groups. These findings can guide farmer decision-making for annual crop and soil management, offering a practical alternative to continuous monoculture that addresses soil degradation and yield decline. The identification of specific microbial taxa associated with beneficial outcomes (e.g., nitrogen-cycling Thaumarchaeota, disease-suppressive Mortierella) provides targets for future research into microbiome-based crop management.