Air-fuel ratio measurement is essential for efficient combustion and emission control in internal combustion engines. This study presents a mathematical model for the voltage response of narrowband zirconia lambda sensors, commonly used in older vehicles. These sensors, while cost-effective, are limited by their narrow operating range and strong temperature sensitivity – especially under rich mixture conditions. Using the Nernst equation, the study models the sensor voltage as a function of oxygen partial pressure. The sensor’s heating element was experimentally characterized, and its resistance–temperature relationship was accurately described using a Rational5 function. A modified form of the Nernst equation enabled voltage-to-AFR conversion across a wide mixture spectrum. Data analysis confirmed that in lean conditions, the voltage response is nearly linear and largely unaffected by temperature, enabling accurate closed-loop control. In contrast, rich mixtures produced highly non-linear and temperature-dependent behavior, making interpretation more complex. To address this, an Arrhenius-based model was successfully applied to the rich-side response, significantly improving accuracy after temperature compensation. The model was implemented and tested on a Ford EEC V ECU with a Ford 302 GT40P engine running on LPG. Modifications to the controller allowed stable closed-loop operation up to lambda = 1.35 using only a narrowband sensor. These results show that with proper modeling and calibration, narrowband lambda sensors can effectively monitor lean mixtures and offer limited utility in rich conditions. This opens new possibilities for retrofitting legacy engine systems without expensive hardware upgrades. Further research is needed to enhance rich-side precision and dynamic temperature correction.