Farmland ecosystems are primary producers of food, feed, fiber, and other natural products. Species diversity constitutes an important foundation in farmland ecosystems, but human activities are greatly accelerating the loss rate of species. Considerable evidence shows that agricultural management threatens biodiversity and negatively affects species richness and abundance of taxa. In major grain-producing areas, reductions in productivity and soil biodiversity and serious pollution problems occur as a result of fertilizer and pesticide use as well as new agricultural technologies. Conservation of biological diversity is considered to be an important strategy to reduce risks to agriculture in the future. Phospholipid fatty acid (PLFA) analysis was employed to examine the effects of fertilization and diversity of weed species on soil microbial community structure in a winter wheat plantation. The experiment used a split-plot design and was established in October 2010. Two fertilization treatments (including chemical fertilizer and organic manure) were applied to the main plots and diversity of weed species (0, 1, 2 and 4 species) were sown in the sub-plots. Wheat was grown in the center of plots and weeds were grown around the wheat plants (all eight plants). The weed species were Avena fatua, Medicago sativa, Cichorium intybus, and Descurainia sophia. For the zero species weed treatments, six plots were grown of wheat plants only. For the 1-species weed treatments, one kind of weed was grown with the wheat in 12 plots. For the 2-species weed treatments, two weed species were grown with the wheat in 12 plots. For the 4-species weed treatments, four weed species were grown with the wheat in six plots. Increased weed diversity significantly increased the soil carbon(C) : nitrogen(N) ratio and pH in both fertilizer treatments, and the C : N ratio was the highest in the 4-species treatment. In the chemical fertilizer treatment, weed diversity significantly affected the fungi:bacteria ratio, which was highest in the 4-species treatment. Fungal and mycorrhizal fungal biomass were lowest in the 1-species treatment (1.0 nmol/g dry soil and 0.4 nmol/g dry soil, respectively), and significantly lower than in the 4-species treatment (1.3 nmol/g dry soil and 0.6 nmol/g dry soil, respectively). In the organic manure treatments, the gram-positive:gram-negative bacterial ratio was lowest in the 0-species treatment compared with the 1-, 2- and 4-species treatments. Mycorrhizal fungal biomass was lowest in the 1-species treatment (1.5 nmol/g dry soil), and significantly lower than in the 4-species treatment (1.8 nmol/g dry soil). In both fertilizer treatments, weed species diversity affected microbial community composition by changing the soil C : N ratio, which was correlated with biomass of various functional groups of soil microbes. Moreover, the shift of the microbial community composition in a different way. Plant species richness is an important factor affecting microbial interactions. Despite the small differences in microbial community structure between different weed diversity treatments, these minor differences will have an important cumulative effect on microbial-driven ecosystem processes. For legumes, Asteraceae, Poaceae, and cruciferous weed species treatments, species specificity had no significant effects on soil microbial biomass and taxa. Thus, species diversity affects microbial community composition. We recommend applying manure to increase soil microbial biomass in farmlands and maintain diversity of weed species. This will lead to changes in microbial community structure to regulate and improve soil ecosystem stability in chemically fertilized farmland. This study is of practical and theoretical significance for 1) our understanding of how plant diversity affects soil microbial community composition and the development of soil ecosystem health, 2) exploring microbial ecological function in maintaining soil ecosystem stability, and 3) revealing plant-soil-microbial interactions and feedback mechanisms.