This brief gives an enhanced result for fuzzy state estimation of discrete-time nonlinear circuits by developing two new technical measures of relaxations. Firstly, a novel balanced matrix method is proposed in which all the involved free matrices are with arbitrary variable orders, and thus a lot of freedom can be introduced into different switching modes for reducing the conservatism caused by existing those free matrices with fixed orders in the recent result. Secondly, an important property is employed in order to further relax the conservatism caused by the mismatch of adjacent normalized fuzzy weighting functions. Therefore, the scope of application of fuzzy state estimation can be enlarged by a large margin and much better estimation performance is obtained than before. Finally, the superiority of the developed approach is tested and verified based on the common benchmark example applied in the literature.