A multirate Kalman filter-based transmultiplexer is proposed in this paper for multirate data transmission systems. The composite effect of combining filters and upsamplers at the transmitter is represented by a multirate state space model. The input signals of the transmultiplexer are contained in the state vector, and the channel noise is modeled as a measurement noise. Additionally, the separation filters and downsamplers at the receiver of a conventional transmultiplexer are replaced by the multirate Kalman filter to eliminate the effect of channel noise and reconstruct the input signals. This design is useful for improving the performance of a transmultiplexer system for simultaneous transmission of many data signals through a single noisy channel. Two simulation examples are provided to illustrate the significant improvement in performance.