Herpes is a usually mild recurrent skin condition in which most infections are unrecognized and undiagnosed. The mechanism of disease is still not well understood. The analysis of peptide sequences of herpes can reveal information which may be useful for understanding the mechanism of disease. In this paper an attempt has been made to develop a model for mining fuzzy amino acid associations in peptide sequences of herpes virus. The uncertainty arising due to variation in length of sequences and this is handled by employing fuzzy sets. Total 9160 sequences were taken from National Centre for Biotechnology Information. After that around 4004 non-redundant peptide sequences of herpes virus filtered to form the dataset. This dataset is trnasformed to fuzzy transaction dataset and their fuzzy support and confidence have been computed. The patterns generated from this model can be useful in understanding the structure, function and interaction of the protein in the disease.