Difference equations with delay are widely used to model the evolution of the biomass of a fish stock (delay difference models). Represented as a state- space model they allow, starting from the data of the annual catches, a relevant Bayesian analysis. For this purpose we can use an hybrid MCMC method combi- ning a Metropolis-Hastings algorithm within a Gibbs sampler, namely the single- component Metropolis-Hastings algorithm.