Global Navigation Satellite System (GNSS) radio occultation (RO) has been an active method to explore the earth’s atmosphere since 1995. RO-inverted bending angles have significantly contributed to atmospheric weather and climate research via data assimilation. However, the used bending angles still contain residual ionospheric error (RIE) after the well-known linear combination correction and should be removed. In this study, we propose to calibrate the RIE using ray-tracing RO bending angle profiles for the first time. The ionospheric background used for ray-tracing was obtained by a Kalman filter data assimilation algorithm through ingesting multisource ionospheric data. More than 380,000 COSMIC RO events during 2008–2013 were simulated via the EGOPS software for statistical calibration. To evaluate the improving performance, we focus on analyzing the deviation of the bending angle from that of National Center for Atmospheric Research climatology results between 60 and 80 km before and after the calibration. It is found that the amplitude of RIE during daytime is significantly reduced. The ‘three troughs’ feature of RIE due to the ionosphere is almost eliminated. The solar activity dependence of RIE can also be calibrated to some extent. We further define a parameter named calibration efficiency to evaluate the calibration performance of the method is > 40% at low latitudes and > 80% at the ‘three troughs’ regions. The results demonstrate that our calibration method works well. It could be potentially used to reduce the RIE in GNSS RO bending angles.