This paper represents an empirical application to investigate the heterogeneity of social demand of agricultural multifunctionality using the Choice Experiment. For this purpose the work focuses on a methodological framework based on a concatenated scheme of different Mixed Conditional Logit and Random Parameters Logit models. This methodological proposal has been implemented empirically to a case study: the agricultural system of cereal steppes in Tierra de Campos in Spain. The results suggest the importance to rely on a joint estimation of different Mixed Conditional Logit models that mitigate the violation of the ceteris paribus condition allowing for a suitable heterogeneity preference analysis for individual attributes.