Fertilization is an important agricultural practice for increasing crop yields. In order to maintain the soil sustainability, it is important to monitor the effects of fertilizer applications on the shifts of soil microorganisms, which control the cycling of many nutrients in the soil. Here, culture-dependent and culture-independent approaches were used to analyze the soil bacterial and fungal quantities and community structure under seven fertilization treatments, including Control, Manure, Return (harvested peanut straw was returned to the plot), and chemical fertilizers of NPK, NP, NK, and PK. The objective of this study was to examine the effects on soil microbial composition and diversity of long-term organic and chemical fertilizer regimes in a Chinese upland red soil. Soil samples were collected from a long-term experiment station at Yingtan (28°15′N, 116°55′E), Jiangxi Province of China. The soil samples (0–20 cm) from four individual plots per treatment were collected. The total numbers of culturable bacteria and fungi were determined as colony forming units (CFUs) and selected colonies were identified on agar plates by dilution plate methods. Moreover, soil DNAs were extracted and bacterial 16S rRNA genes and fungal 18S rRNA genes were polymerase chain reaction amplified, and then analyzed by denaturing gradient gel electrophoresis (DGGE), cloning, and sequencing. The organic fertilizers, especially manure, induced the least culturable bacterial CFUs, but the highest bacterial diversity ascertained by DGGE banding patterns. Chemical fertilizers, on the other hand, had less effect on the bacterial composition and diversity, with the NK treatment having the lowest CFUs. For the fungal community, the manure treatment had the largest CFUs but much fewer DGGE bands, also with the NK treatment having the lowest CFUs. The conventional identification of representative bacterial and fungal genera showed that long-term fertilization treatments resulted in differences in soil microbial composition and diversity. In particular, 42.4% of the identified bacterial isolates were classified into members of Arthrobacter. For fungi, Aspergillus, Penicillium, and Mucor were the most prevalent three genera, which accounted for 46.6% of the total identified fungi. The long-term fertilization treatments resulted in different bacterial and fungal compositions ascertained by the culture-dependent and also the culture-independent approaches. It was evident that more representative fungal genera appeared in organic treatments than other treatments, indicating that culturable fungi were more sensitive to organic than to chemical fertilizers. A very notable finding was that fungal CFUs appeared maximal in organic manure treatments. This was quite different from the bacterial CFUs in the manure, indicating that bacteria and fungi responded differently to the fertilization. Similar to bacteria, the minimum fungal CFUs were also observed in the NK treatment. This result provided evidence that phosphorus could be a key factor for microorganisms in the soil. Thus, despite the fact that culture-dependent techniques are not ideal for studies of the composition of natural microbial communities when used alone, they provide one of the more useful means of understanding the growth habit, development, and potential function of microorganisms from soil habitats. A combination of culture-dependent and culture-independent approaches is likely to reveal more complete information regarding the composition of soil microbial communities. Long-term fertilization had great effects on the soil bacterial and fungal communities. Organic fertilizer applications induced the least culturable bacterial CFUs but the highest bacterial diversity, while chemical fertilizer applications had less impact on soil bacterial community. The largest fungal CFUs were obtained, but much lower diversity was detected in the manure treatment. The lowest bacterial and also fungal CFUs were observed in the NK treatment. The long-term fertilization treatments resulted in different bacterial and fungal compositions ascertained by the culture-dependent and also the culture-independent approaches. Phosphorus fertilizer could be considered as a key factor to control the microbial CFUs and diversity in this Chinese upland red soil. Soil fungi seem to be a more sensitive indicator of soil fertility than soil bacteria. Since the major limitation of molecular methods in soil microbial studies is the lack of discrimination between the living and dead, or active and dormant microorganisms, both culture-dependent and culture-independent methods should be used to appropriately characterize soil microbial diversity.