Pulsars rotate extremely stably, they can be used as time standard. However, pulsar time defined by single pulsar is influenced by several noise resources. To weaken these influences to get a more stable time scale, one can take ensemble analysis to obtain the ensemble pulsar time. Since different pulsars have different rotation frequency, and the noise source have different influence on different frequency of each pulsar，we can apply wavelet analysis to integrate the pulsar time, and this algorithm can take consideration of the long-term and short-term stability of pulsar time. On the other hand, the pulsar timing residuals are caused by reference atomic clock and pulsar itself, so we can apply Wiener filtration analysis to distinguish the two parts, and then calculate the ensemble pulsar time with the residuals after removing the error of reference atomic clock. Based on both of the two algorithms, a new algorithm of Wiener filtration in wavelet domain has been presented, which combines the merits of wavelet analysis and Wiener filtration and can remove the noise influence to pulsar time more effectively. Experimental results prove that Wiener filtration in wavelet domain can reduce noise influence of pulsar timing residuals effectively.