The objective of this study was to compare the automated ribosomal intergenic spacer analysis (ARISA) and the denaturing gradient gel electrophoresis (DGGE) techniques for analysing the effects of diet on diversity in bacterial pellets isolated from the liquid (liquid-associated bacteria (LAB)) and solid (solid-associated bacteria (SAB)) phase of the rumen. The four experimental diets contained forage to concentrate ratios of 70:30 or 30:70 and had either alfalfa hay or grass hay as forage. Four rumen-fistulated animals (two sheep and two goats) received the diets in a Latin square design. Bacterial pellets (LAB and SAB) were isolated at 2 h post-feeding for DNA extraction and analysed by ARISA and DGGE. The number of peaks in individual samples ranged from 48 to 99 for LAB and from 41 to 95 for SAB with ARISA, and values of DGGE-bands ranged from 27 to 50 for LAB and from 18 to 45 for SAB. The LAB samples from high concentrate-fed animals tended (p < 0.10) to show greater peak numbers and Shannon index values than those isolated from high forage-fed animals with ARISA, but no differences were identified with DGGE. The SAB samples from high concentrate-fed animals had lower (p < 0.05) peak numbers and Shannon index values than those from animals fed high-forage diets with ARISA, but only a trend was noticed for these parameters with DGGE (p < 0.10). The ARISA detected that animals fed alfalfa hay diets showed lower (p < 0.05) SAB diversity than those fed grass hay diets, but no differences were observed with DGGE. No effect of forage type on LAB diversity was detected by any technique. In this study, ARISA detected some changes in ruminal bacterial communities that were not detected by DGGE, and therefore ARISA was considered more appropriate for assessing bacterial diversity of ruminal bacterial pellets. The results highlight the impact of the fingerprinting technique used to draw conclusions on dietary factors affecting bacterial diversity in ruminal bacterial pellets.