A two-stage biomass model is developed in the Bayesian framework that allows us to assimilate various sources of information. A method that makes use of ancillary length frequency data is developed to provide an informative prior distribution for the intrinsic biomass growth rate parameter and its annual variability. The new Bayesian model provides substantial improvement to the existing stock assessment method used by ICES. Considering a time-varying g parameter improves model fit and improves the ecological realisms of the model according to the sensitivity of the cuttlefish population dynamics to environmental fluctuations. We present results of the English Channel cuttlefish stock assessment updated with the new Bayesian model. The model also provides predictions of the unexploited biomass in winter based on survey data, and help managing the stock in case of strong depletion.