This thesis presents a comprehensive study on emissions predictive control modelling for hybrid electric scooters. Two approaches were investigated on a constructed hybrid electric scooter. The first approach involves developing a hybrid electric scooter dynamic model using MATLAB-Simulink and the second involves the development of an Emissions Predictive Model using artificial neural network. The hybrid electric scooter model was developed to further understand and analyze as well as to predict its performance and emissions before proper construction of the prototype begins. The MATLAB-Simulink model consists of four integrated models that formed the complete hybrid scooter model: Battery Model, Engine Model, DC Motor Model and the Vehicle Dynamics Model. The multi-mode controller predicts the required parameters to operate the scooter in an optimize condition. Experimental data were gathered and thus compared to the simulated data to check the model's feasibility and accuracy on four distinct driving cycles: Modified Urban Dynamometer Driving Schedule, New York City Cycle, European Driving Cycle and the Modified Highway Fuel Economy Driving Schedule. Results showed that the developed multi-state hybrid electric scooter model was accurate and feasible with predictive errors of ±10 % for emission levels and fuel economy on the European Driving Cycle. Simulated results were also compared to the existing literature and it was found that the qualitative trends were similar. By having a high-confidence simulation model, performance of the hybrid electric scooter were also simulated over the mentioned driving cycles demonstrating the optimization strategy of the multi-state control system. For the second approach, the Emissions Predictive Model was then built using artificial neural network techniques to predict the following tailpipe emissions gases; CO, CO2, HC and O2. Three feed-forward neural network models were investigated and compared in this study; back-propagation, optimization layer-by-layer and radial basis function networks. Based on the experimental setup, the neural network models were trained and tested to accurately predict the effect of the engine operating conditions on the emissions by varying the number of hidden nodes. The selected optimization layer-by-layer network proved to be the most accurate and reliable predictive tool with prediction errors of ±5 %. The effect of the engine operating conditions on the tailpipe emissions for a scooter is shown to display similar qualitative and quantitative trends between the simulated and the experimental data. Having accurate predictive models for emissions and fuel economy enable the hybrid electric scooter to be optimized via modelling and simulation before proper construction begins. The developed emissions predictive models could act as a virtual emissions sensor replacing costly hardware for the developed physical hybrid electric scooter. This study provides a better understanding in effects of engine process parameters on tailpipe emissions for the hybrid electric scooter as well as for general hybrid vehicular applications