Challenging limits for NOx in BS VI emission legislation demand high performance conversion techniques. Exhaust after treatment systems such as Lean NOx Trap and selective catalytic reduction can provide effective reduction of Engine out NOx emissions. From the moment of engine start, these systems require reliable signal input of the NOx sensor. The efficient use of such DeNOx systems demands earliest possible activation of the upstream NOx sensor. However, attainment of the sensor dew point delays reliable measurements. Data collected from the emission test cycles, WLTC and NEDC indicates the unavailability of NOx sensor from the beginning of the test cycle. Hence, requirement of a NOx model is inevitable to estimate NOx emissions till sensor reaches its dew point. The scope of this paper is to elucidate the calibration process through which a robust estimation of the NOx emissions can be made in different engine modes along with varying ambient conditions. The NOx model is capable of estimating the real time NOx emission using a map based approach taking in to account multiple factors such as coolant temperature, ambient air temperature and pressure. Initially, a set of reference maps is generated for the required engine modes by performing mappings at the dynamometer. With state of the art climatic dynamometers, calibration tools and techniques, it is possible to introduce a robust calibration of the model which the De NOx system can rely upon. This paper contains data collected from measurements at emission, non-emission dynamometers, cold and hot temperature test trips, altitude test trips. The validated NOx model presented in this paper is able to predict real time NOx emission with a deviation of less than 5%.