Adaptive echo cancellers are currently being studied for application to hands-free telephone sets with high speech quality. The major problem is the acoustic echo control from the loudspeaker to the microphone. This proposal presents a new concept by using linear prediction techniques in combination with a conventional adaptive echo canceller. Computer simulations show that the introduced algorithm improves significantly the performance compared to conventional methods. In this paper the new structure is described. A comparison of the results with the state of the art indicates that the complexity can be reduced by one third.