Electrochemical sensors integrated with CMOS circuits play a pivotal role in quantifying various chemicals across a huge spectrum of applications, prominently in distributed diagnostics and remote monitoring. When designing integrated devices, meticulous consideration of the electrochemical sensor model becomes imperative, particularly in the CMOS design of electrochemical front-ends. This paper proposes a thorough analysis of the prevalent equivalent circuits for the sensor often considered by CMOS designers, elucidating the substantial limitations inherent in two models largely used and widely recognized in literature. On the other hand, a more comprehensive and precise electrical model is proposed, poised to enhance the robustness of the resulting circuit designs. The comparison between in-vitro measurements and simulation results by considering these three models demonstrates the inherent inaccuracies in the two models conventionally used, while highlighting the improved accuracy of the proposed new electrical model. Using the dopamine detection with a fabricated digital potentiostat as a model example of CMOS design, the conventional electrical equivalent circuits manifest a noteworthy 60% difference in current readings within a sampling time of 600 ms, compared to the measurements, which are instead better followed by the more robust proposed new model. Furthermore, when evaluating saturation risk, the conventional models exhibit a staggering 110% error compared to the measurements and the proposed model.