For a better understanding of long timescale transport dynamics, the rescaled range analysis techniques, the autocorrelation function (ACF) and the probability distribution function (PDF) are used to investigate long-range dependences in edge plasma fluctuations in an HT-6M tokamak. The results reveal the self-similar characters of the electrostatic fluctuations with self-similarity parameters (Hurst exponent) ranging from 0.64 to 0.79, taking into consideration the E-r ×B- rotation-sheared effect. Fluctuation ACFs of both the ion saturation current and the floating potential, as well as PDF of the turbulence-induced particle flux, have two distinct timescales. One corresponds to the decorrelation timescale of local fluctuations (μs) and the other lasts to the order of the confinement time (ms). All these experimental results suggest that some of the mechanisms of the underlying turbulence are consistent with plasma transport as characterized by self-organized criticality (SOC).