The purpose of this study was to quantify the short and long-term correlation characteristics of heart rate variability with atrial fibrillation (AF). Long term analysis was performed on the entire 24-hour heart rate series by applying two independent methods: a) analysis of correlation dimension, and b) analysis of fractal scaling properties using detrended fluctuation analysis (DFA). Short term analysis was performed on non-overlapping series of 500-beat segments, spanning the entire 24 hours, by applying two complementary techniques: the DFA analysis and the autocorrelation function (ACF). We obtained consistent results among the subjects for both long and short-term analysis. DFA analysis of the 24-hours showed a two-slope “cross-over” behavior with the inflection point occurring at about 100 beats. The average slope was αl = 0.55±0.05 (below 100 beats) and α2 = 1.16±0.06 (above 100 beats). DFA analysis on the short 500-beat segments showed a single slope behavior in all subjects (α = 0.56±0.04), with the exception of a patient who underwent electrical cardioversion. The second point of the ACF follows an almost identical behavior (α = -0.49±0.03). Therefore, for time scales less than 100 beats the local beat-to-beat behavior is reminiscent of a “white noise” (uncorrelated) process. In contrast, for time scales greater than 100 beats, the heart rate scaling properties in AF are surprisingly comparable to those in sinus rhythm (long-range Organization). However, we find no evidence of low-dimensional deterministic chaos in AF.