In this research article, a new finite impulse response (FIR) filter for discrete time Takagi-Sugeno (T-S) fuzzy system is proposed to estimate the state of non-linear system. Compared with conventional T-S fuzzy filters which are known to have infinite impulse response (IIR) and require whole data from initial time to the current time to operate, the proposed filter is a FIR filter which guarantees to have robustness to unknown initial data. In addition, the proposed FIR filter satisfies the unbiased condition with the minimum variance error. Simulation result demonstrates that the proposed FIR filter is more robust and fast in convergence under the incorrect initial data than the existing non-linear filters with the same condition by using tunnel diode model.