Abstract This paper forms Part II of a two-part series which describes the overall construction of the computer simulation and the estimation algorithm of a static-state estimator for a power station boiler. The estimator developed is based on the mathematical model established in Part I of this series. Analysis of the observability of the measurement equations shows that the estimator is observable. The estimator, based on the use of the weighted least square approach, has been tested in a random noise environment. Several experiments on filtering out normal distributed noises with different magnitudes of the variance have been performed and the numerical results show that the estimator works well. The J-index test and R w and R n tests have been applied to the processes of detection and identification of bad data respectively, and a comparison between these testing criteria has led to the preference of a combination of the J-index and R n tests in the bad data processing. The ability of the estimator to process multiple bad data is also tested and discussed.