Cardiovascular disease (CVD) is one of the leading causes of death worldwide, emphasizing the critical need for frequent heart monitoring to detect abnormalities in heart rhythm. While current electrocardiogram (ECG) devices mostly acquire and display ECG signals for clinician’s manual interpretation, they lack the capability for self-classification of multiple complex arrhythmias. To address these issues, previous in-house work has developed a 2-stage arrhythmia classification technique, which the first stage consists of two distinct beat arrhythmia classifiers and rhythm arrhythmia classifiers, respectively, whereas the second stage acts as the final decision maker of arrhythmia detection based on the detection result from the first-stage classifiers. However, these algorithms show some inaccuracies in detecting certain arrhythmias especially when two algorithms are integrated together. This study aims to enhance the accuracy of the second stage of the existing two-stage arrhythmia classification algorithm by applying a decision tree approach. In this study, Olimex EKG/EMG Shield and Arduino Uno board are also utilized to design a ECG signals acquisition unit, to acquire signals from the Fluke ProSim 3 vital sign simulator for offline ECG signals recordings for system processing and functionality verification. Results reveal that the overall accuracy of the arrhythmia detection is 90.83% based on the proposed enhanced algorithm.