We observed a remarkable near-to-unity correlation between the time series of Acoustic Emissions (AEs) collected at two stations approximately 300 km apart from each other and located along the Apennine belt (Italy). This finding prompted us to verify the hypothesis that AE signals can carry with them an indication of anomalies in a crustal stress trend, possibly related to earthquake occurrences. Thus, we checked the ability of Recurrence Quantification Analysis and Fractal Analysis as applied to AE to identify signal phase transitions before the crisis occurs. The sharp drop of the Percent of Determinism after its maximum value, and simultaneously with minimum values of the Fractal Dimension (D), few days before some seismic events take place, seems to point to the relevance of the proposed approach as precursor detection.