Denaturing gradient gel electrophoresis (DGGE) is one of the most frequently used methods for analysis of soil microbial community structure. Unbiased PCR amplification of target DNA templates is crucial for efficient detection of multiple microbial populations mixed in soil. In this study, DGGE profiles were compared using different pairs of primers targeting different hypervariable regions of thirteen representative soil bacteria and clones. The primer set (1070f-1392r) for the E. coli numbering 1,071-1,391 region could not resolve all the 16S rDNA fragments of the representative bacteria and clones, and moreover, yielded spurious bands in DGGE profiles. For the E. coli numbering 353-514 region, various forward primers were designed to investigate the efficiency of PCR amplification. A degenerate forward primer (F357IW) often yielded multiple bands for a certain single 16S rDNA fragment in DGGE analysis, whereas nondegenerate primers (338f, F338T2, F338I2) differentially amplified each of the fragments in the mixture according to the position and the number of primer-template mismatches. A forward primer (F352T) designed to have one internal mismatch commonly with all the thirteen 16S rDNA fragments efficiently produced and separated all the target DNA bands with similar intensities in the DGGE profiles. This primer set F352T-519r consistently yielded the best DGGE banding profiles when tested with various soil samples. Touchdown PCR intensified the uneven amplification, and lowering the annealing temperature had no significant effect on the DGGE profiles. These results showed that PCR amplification bias could be much improved by properly designing primers for use in fingerprinting soil bacterial communities with the DGGE technique.