To gain knowledge on the precision of an in vitro method for characterisation of the fermentability of dietary fibres, this study aimed to evaluate the repeatability and reproducibility of such a method. Substrates used were citrus pectin (CP), fructo-oligosaccharides (FOS), guar gum (GG), sugar beet pulp (SBP) and wheat middlings (WM). Each substrate was incubated with faecal inoculum from five cats with three replicates for each substrate–cat combination. Gas production was measured continuously during the 48 h incubation and SCFA and organic matter disappearance (only SBP and WM) were determined after incubation. Four consecutive runs were performed. The within-run variability (repeatability) was generally lower for the more simple and pure substrates (CP, FOS, GG) than for the more complex substrates containing mixtures of fibres (SBP, WM). Replicates showed high variability, in particular for SCFA profiles and parameters of gas production kinetics. The between-run CV (reproducibility) for the measured parameters were, in general, below 10 % for CP, FOS and GG and higher values were obtained for SBP and WM. It is concluded that for precise dietary fibre characterisation, the number of replicates should be multiple and adjusted according to the variability of the parameters of interest and the complexity of fibres. The method yielded reproducible results with some variation in absolute values obtained, which may have an impact on the significance level of the differences among substrates.