A novel approach for the nonlinear characterization of Electrocardiogram (ECG) signals has been developed. The new developed methodology is based on a numerical algorithm that extracts the value of dinfinity (d-infinite) characterizing the asymptotic chaotic behavior of a system. This algorithm also extracts a measure of the maximum Lyapunov exponent and it is applicable to time series where the knowledge of the system structure and laws is not necessary. In order to prove the significance of the extracted parameters, the presented algorithm was applied on a statistically significant number of ECG signals taken from the MIT-BIH database and including normal subjects and subjects affected by arrhythmia and ventricular arrhythmia. A systematic study, analyzing how dinfinity varies with initial condition was performed showing the sensitivity of such parameter to the initial conditions. Furthermore, two maps, one presenting the maximum Lyapunov exponent and the other the dinfinity versus a control parameter II, as a measure of the rate variation, were drawn using the parameters extracted by the experimental data. They clearly show three distinguishable zones where the normal subjects and the subjects affected by the two different pathologies can be mapped and discriminated. Concluding, the newly presented algorithm, thanks to its implementation features and its effectiveness, it lends itself to future real-time implementation for clinical application in the early diagnosis of cardiac pathologies.