Ionospheric correction, using a linear combination of the dual frequency observations from the occulting GPS satellite, is required to obtain high quality GPS radio occultation (RO) profiles of bending, refractivity, temperature, pressure and humidity in the neutral atmosphere. However, the ionospheric correction that is routinely applied in GPS RO data processing is not perfect. Small errors induced by the ionosphere remain in the so-called “ionospheric corrected” bending angles and can affect neutral atmospheric profiles at heights above 25 km. These errors have a random component which must be addressed for weather forecasting applications. Smaller, but more problematic is a systematic bias that can be caused by residual ionospheric errors. We have found experimentally that this bias changes with solar activity and that it is correlated with the 10.7 cm solar flux. Thus this ionosphere-induced bias, if left uncorrected, can introduce a false neutral atmospheric signal into the RO data record. Because RO data are believed to be particularly useful for stable monitoring of long-term climate trends it is important to understand and correct the residual ionospheric bias that changes between night and day and with the 11-year solar cycle. This presentation discusses how we can correct the RO data climate record from COSMIC and other missions to minimize the ionospheric contamination.