In recent years there has been a renewal of interest for fuzzy associative memories. However it is difficult to apply fuzzy associative memories to complex knowledge processing, because associative memories have poor storage capacity. In this study, we propose a method to resolve associative memories' perplexity which is due to the poor storage capacity. The term `associative memory's perplexity' can be defined by the fact that, after the recalling processing, several undesired nodes may have almost the same activation as the desired ones. Furthermore we apply the proposed method to the construction of a facial expression model and show its effect.