This paper describes an approach to identify and change the measurement weights used in Weight Least Square (WLS) estimation method employed in State Estimation (SE). In practice, the individual measurement is assigned with their own weight factor based on technical experience by the engineers. However, uncertainty in analog measurements could occur in a real time system. Therefore, those measurements that are assigned with high weighting factor need not be a good data. Method of Autoregressive (AR) Burg is proposed in this paper to identify and change the measurement weights of the bad measurements if the related measurement is assigned with high weight factors. Simulated measurement data and state estimation results are used as historical data. The proposed method is tested on IEEE 14 bus network and SE provides reliable estimation.