The multivariable autoregressive (AR) model identification technique has been applied in the study of the boiling water reactor core stability test analysis. It has been demonstrated through the analysis of core stability tests performed at the Peach Bottom-2 reactor, so that the AR model technique is effective in estimating core stability performance. Neutron flux to dome pressure open-loop stability performance is estimated by two methods, the ordinary correlation method and the AR model technique. Results obtained by both methods are in good agreement. The AR model technique can provide closed-loop decay ratios. This kind of decay ratio is considered to represent the actual core stability characteristic. Based on these test analysis results, the closed-loop in-reactor characteristic is more stable than the open-loop characteristic, which is usually considered to be the stability index for the reactor core. It was attempted to evaluate error in the AR model technique through indirect ways. It has been concluded that the AR model technique for the stability test data analysis is quantitatively highly effective in identifying and evaluating the core stability characteristics.