Accurate in-cylinder air charge estimation is important for engine torque determination, controlling air-to-fuel ratio, and ensuring high after-treatment efficiency. Spark ignition (SI) engine technologies like variable valve timing (VVT) and exhaust gas recirculation (EGR) are applied to improve fuel economy and reduce pollutant emissions, but they increase the complexity of air charge estimation. Increased air-path complexity drives the need for cost effective solutions that produce high air mass prediction accuracy while minimizing sensor cost, computational effort, and calibration time. A large number of air charge estimation techniques have been developed using a range of sensors sets combined with empirical and/or physics-based models. This paper provides a technical review of research in this area, focused on SI engines. The purpose is to provide an overview of current SI engine air charge estimation techniques and their performance in key areas such as transient and steady-state accuracy, calibration effort and computational load. Several common air estimation methods are replicated and compared over similar operating conditions. Particular focus is given to methods utilizing mass air flow (MAF) sensors, speed-density algorithms, and observers. Speed density approaches evaluated include those with neural networks and physics-based volumetric efficiency models. Observer methods employing open-loop air charge, high gain input and Extended Kalman Filters (EKF) are also evaluated and compared.