Considering the highly nonlinear and non stationary features of the ECG signal proven by latest researches, the most appropriate methods of analysis are based also on the nonlinear dynamics; we used a modified root mean square analysis of a random walk, named detrended fluctuations analysis (DFA), which proved efficacy as diagnostic tool and advantages on the existing related methods.Our studies were conducted on two groups of young and old subjects, totally 24 patients, and the different behavior being related to the underlying dynamics of the heartbeat.We compare the data results using the DFA method: first when a line segment is fitted and second, when a quadratic polynomial is fitted in the least-squares sense Applications of this analysis may lead to new and safe diagnosis for patients and to evaluation of the patient status in systemic diseases that may affect in time the cardiovascular system.