There are many different serious cardiac arrhythmias. To detect cardiac arrhythmia of a patient, the present invention proposes to detect RR intervals of the patient wherein each RR interval is an interval between two heart beats, and to simulate standard probability density histograms of ΔRRs by means of a suitable probability distribution calculated through at least one mathematical formulae, wherein ΔRR is a difference between two successive RR Intervals. Test probability density histograms of ΔRRs of the patient are constructed from the detected RR intervals. Finally, the standard and test histograms are compared to detect whether the patient suffers from cardiac arrhythmia. As non limiting examples, the standard probability density histograms of ΔRRs are modeled by a mathematical equation selected from the group consisting of : the Lorentzlan distribution, the Gaussian distribution, the Student's t-distribution, and a probability distribution including a linear combination of the Lorentzian and Gaussian distribution.