This paper presents an adaptive, heart-model-based ECG compression method. After a conventional pre-filtering the waves from the signal are localized and the model's parameters are determined. The structure of the algorithm allows real-time adaptation to the heart's state. The compression, for better comparison, was performed for one and more channels from the MIT/BIH database samples. The compression ratio depends on the maximal allowed root mean square reconstruction error (RMSRE). As a second classification criterion we applied the performance of the signal detection method from the compacted data. We used an adaptive entropy encoder to reduce the redundancy. The major advantage of this method is the possibility to accomplish a real-time, adaptive and patient specific encoding with relatively low computational power, ideal for telemetry measurements. This research is supported by the Hungarian Foundation for Scientific Research, Grant T29830 and FKFP0301/0999 Project.