Using the modern time series analysis method, based on the autoregressive moving average (ARMA) innovationmodel, the unified fast suboptimal fixed-interval white noise smoothers is presented by the truncation of polynomial matrices forsystems with correlated noises. They can obviously reduce the computational burden, and can be expresses as the innovationfilter and Wiener filter. The formulas of truncated error and truncated index are given. A simulation example for Bemoulli-Gaussian white noise shows its effectiveness.