In this paper a three-layered feedforward self-organizing neural network model is proposed based on interval-valued fuzzy logic neurons in order to realize the interval-t

alued fuzzy C-means （IVFCM） clustering analysis. The inputs/outputs of the first and second layer neurons and weights between them belong to foe interval-valued fuzzy sets I [0

1]. The neurons in the first layer are intervalvalued linear neurons; The neurons in the second layer are interval-valued fuzzy equality neurons （IVFEN） which realize the matching computation between the input samples and all of the clusters. Interval-valued fuzzy equality relation is used to compute matching. In order to define the model of IVFEN

the computation method of interval-valued fuzzy equality relation is deduced based on point-valued fuzzy equality relation. The neurons in the third layer are fuzzy competition neurons. The outputs of the third-layer neurons represent the fuzzy classification result of the input sample. Furthermore

intervalvalued fuzzy competition learning （IVFCL） is proposed to train the IVFCM neural network （IVFCMNN）