Flexible fuel vehicles (FFVs) are able to operate on a blend of ethanol and gasoline in any volumetric concentration of up to 85% ethanol (93% in Brazil). The estimation of ethanol content is crucial for optimized and robust performance in such vehicles. Even if an ethanol sensor is utilized, an estimation scheme independent of the ethanol sensor measurement retains advantages in enhancing the reliability of ethanol estimation and allowing on-board diagnostics. It is well-known that an exhaust gas oxygen (EGO) sensor could be utilized to estimate the ethanol content, which exploits the difference in stoichiometric air-to-fuel ratio (SAFR) between ethanol (9.0) and gasoline (14.6). The SAFR-based ethanol estimation has been shown to be prone to large errors with mass air flow sensor bias and/or fuel injector shift. In this paper, an ethanol estimation scheme is proposed by additionally using measurement of in-cylinder pressure, which essentially exploits the difference in latent heat of vaporization (LHV) between ethanol and gasoline, and hence detects the charge cooling effects of ethanol. Due to the additional and independent information introduced by the new algorithm, the proposed ethanol content estimation is more tolerant to air flow sensor drifts and injector shifts.