Abstract—A Physical model for the long-term deterministic behavior associated with instability (or drift) in pH-sensitive ion-selective field effect transistors (ISFET’s) has been employed to distinguish between drift and the inherent low-frequency (1/f) noise, both of which corrupt the sensor output signal. Based on this model, a physically-informed method for sensor signal enhancement in the time-domain has been proposed, which relies on proper windowing of the sensor output. The random-like evolution in time associated with a weak power-law relaxation towards equilibrium is the key aspect of the model for drift that is utilized in the proposed corrective scheme. Drift may be interpreted as a weakly chaotic behavior, exhibiting anomalous dynamics characterized by ageing, which reflects a weak (power-law) relaxation towards equilibrium.