Abstract: The consistency analysis of the signal feature is an important method to estimate the relationship between the event and the signal in the detection of seismic event. The consistency of amplitude can be used for determination of association through the comparison of amplitude residuals with numerous signal features on time and frequency domain. Different phases comply with the different propagation and attenuation law of amplitude-distance-depth. In this paper, the distribution of sample size varying with the distance of teleseism and ultra-teleseism phases such as P, PcP, PKP, PKPab and PKPbc based on the IDC is made. The model of amplitude-magnitude- distance-depth is constructed by iterative regression based on the residual statistics. The model is estimated through the standard deviation and mean of residual of magnitude. The range of standard deviation of residual of magnitude is 0.30−0.36. It could satisfy the requirements of phase consistency. Besides the mean of residual of magnitude of PKPbc phase is greater than 0.03, the rest phases is about 0.01. It demonstrates that the system deviation of the model is so tiny that it could be neglected. The application research of model is carried out by three different cases such as comparison of predicted amplitude, detection of magnitude consistency and magnitude calculation of ultra-teleseism events. The application research demonstrates that these models could be applied to routine seismic monitoring.