This paper presents an exhaust gas recirculation (EGR) rate estimation algorithm for a turbocharged diesel engine. Accurate estimation of the EGR rate is important for precise air system control of a diesel engine. In order to estimate the EGR rate accurately, we developed an adaptive observer using a model reference identification scheme (MRIS). A linear parameter varying model for the intake manifold pressure dynamics is derived as the reference model for the adaptive observer. The intake and exhaust temperature models are developed through an empirical approach. The MRIS is used to design an update rule for the adaptive observer. Convergence of the proposed observer is proven by using the Lyapunov stability criterion. The proposed observer is implemented in a real-time embedded system and validated via engine experiments.