Recognizing human emotions from partial facial features is quite hard to achieve reasonable accuracy. In this paper, we propose to use a tree structure representation to simulate as human perceiving the real human face and both the entities and relationship could contribute to the facial expression features. Moreover, a new structural connectionist architecture based on a probabilistic approach to adaptive processing of data structures is presented to generalize the FacE Emotion Tree Structures (FEETS). We demonstrated the robustness of our proposed system in recognizing the correct emotion based on partial face features. The system yields an accuracy of about 90% for subjects with partial face covered by artifacts.