In this study, we use the cross-impact analysis to define the relationship among impact, mutation, and outbreak of bird flu. Then we use the distribution rank, which is developed by us over last several years, to quantify the mutations from amino acid sequences of 134 hemagglutinins and 97 neuraminidases. With the help of Bayesian equation, we calculate the probability of occurring of mutation in H5, H6, and H9 hemagglutmins, and N1 and N2 neuraminidases. Finally, we estimate the probability of occurring of mutation with different intensities of an impact. Although we have no means to predict an impact, which is severe enough to lead to the mutations in hemagglutinins and neuraminidases resulting in the Outbreak of bird flu, we can in principle monitor the changes in distribution rank along the time course, and predict the trend of mutations, even to predict the degree of outbreak of bird flu. (C) 2004 Elsevier Inc. All rights reserved.