Valuing environmental goods through benefit transfer is based on information from the set of past empirical studies of similar or identical sites. This paper proposes a Bayesian approach to model prior information on mean consumer surplus. The prior distribution is defined as a convex combination of the results from past empirical studies. An advantage of the Bayesian approach is that the prior distribution could be combined with on-site sample information in order to improve predictions on the benefits of the policy site. The model is applied to data from national parks in Spain. The results show that the benefits of policy sites are more sensitive to the prior information for smaller sample sizes. This implies that the prior distribution can be seen as a substitute for on-site sample information.