Solar wind plasma as a cause of radio signal delay plays an important role in solar and planetary science. Early experiments studying the distribution of electrons near the Sun from spacecraft ranging measurements were designed so that the radio signal was passing close to the Sun. At present, processing of spacecraft tracking observations serves a different goal: precise (at metre level) determination of orbits of planets, most importantly Mars. The solar wind adds a time-varying delay to those observations, which is, in this case, unwanted and must be subtracted prior to putting the data into the planetary solution. Present planetary ephemerides calculate the delay assuming a symmetric stationary power-law model for the solar wind density. The present work, based on a custom variant of the EPM lunar–planetary ephemeris, questions the accuracy and correctness of that assumption and examines alternative models based on in situ solar wind density data provided by OMNI and on the ENLIL numerical model of the solar wind.