Automate the classification of agro-industrial products, requieres solving problems such as the subjective interpretation of parameters and characteristics that define the selection process. Evaluate these parameters is subjective, as both consumers and producers estimate different values depending on their interests, with reference to the offer and demand of products in the market. Likewise, the selection criteria for these parameters define individual criteria and / or group, making the process as automated sorting takes place in separate production lines, without considering the characteristics of products together. This paper proposes a fuzzy model of dynamic membership functions that vary depending on external factors (interest), seeking to automate the classification of agro-industrial products under the detailed factors. 500 were classified Peruvian Scallops, 100 for each kind, achieving an efficiency of 91 %, resulting in a recommended model for similar problems. © 2012 IEEE.