A real-time ECG beat detection and classification system integrated with a decimation filter is presented in this paper. The decimation filter with a decimation ratio of 128 is utilized with four stages to optimize power. The beat detection algorithm is based on multi-scale Haar wavelet transform. The first step of the beat detection algorithm is to define a region composed of 128 sample points. The next step is to define the beat position. The beat detection algorithm which is verified with data from MIT/BIH Arrhythmia Database achieves a sensitivity of 99.67% and a positive predictability of 99.59%. The used heartbeat classification algorithm is called maximum similarity comparisons. The parameter of the classification is programmable for doctors. This proposed system can realize multiple functions in real time. This chip is fabricated in a 0.18 μm standard CMOS technology with chip area of 4.84 mm2, which achieves low power consumptions and can carry out a long term cardiac monitoring.