The in vitro cumulative gas production technique can be used to assess microbial activity of a complex community, in relation to fermentation of a particular energy source. Therefore, in combination with an in vivo study to examine the effects of two different diets for weaning piglets, microbial activities of faeces were compared from animals on the two different diets. The two diets were: CHO diet [containing added fermentable carbohydrates, including sugarbeet pulp (SBP) and wheat starch (WST)], and control diet without any added fermentable carbohydrates. Neither diet contained antibiotics or extra added copper. Twenty‐four piglets were selected from 12 litters (two per litter), weaned at 4 weeks of age (neither creep feeding nor any antibiotic treatment before and during the study), and introduced to one of the two diets. After 9 days on the diet, faecal samples were collected from selected animals, and tested for their activity in terms of gas production kinetics, and end‐products such as volatile fatty acids, ammonia and dry matter disappearance of the two test substrates SBP and WST. The bacterial diversity was also analysed before and after in vitro fermentation using denaturing gradient gel electrophoresis analysis of amplified 16S rRNA genes. There were differences both in kinetics and end‐products of the substrates. More interestingly, significant differences were detected between inocula, although mainly in terms of fermentation kinetics of the two substrates. With the CHO inoculum, SBP was fermented faster than with the control, while this effect was reversed for WST. Significantly higher diversity, as measured by DGGE fingerprint analysis, was detected in the microbial community enrichment on SBP as compared with WST at the end of fermentation. The difference between the kinetics of SBP compared with WST fermentation by faecal microbiota from the CHO diet fed piglets suggests better adaptation to SBP fermentation than to WST fermentation. The WST fermentation was more unexpected, given that a significant amount of starch is known to be fermentable by the small intestinal microbiota. It was concluded that the microbial community composition and activity in the GIT may be changed in response to diet, and that this change can be detected in vitro. Copyright © 2005 Society of Chemical Industry