Abstract: A modeling method is proposed to establish T-S fuzzy system based on immune theory.Using the grid-type method to partition the space,regarding the modeling data as the antigen and the fuzzy rules as the antibody,this method determines the structure of T-S system with immune theory and adopts the least square method to estimate consequent parameters of the linear rule.With many advantages such as simplicity,fast learning and powerful realtime performance,this method is suitable for online learning and struture adjustment of multi-input fuzzy systems.