Neurological evidence was uncovered by A. Demasio revealing the existence of 'emotional intelligence' and its importance. Following this breakthrough many computational models of emotions were developed. Although, psychological research on emotions recognized memory and experience as the main factors that define and shape the complexity of the emotional process, existing computational models of emotions did not incorporate experience or learning. We are proposing a model of an agent named PETEEI - a PET with Evolving Emotional Intelligence. PETEEI was modeled to produce emotions according to its own experience. FUrthermore, PETEEI filters and expresses emotions according to its own moods and previous experience.