Charge exchange between solar wind protons and neutral atmospheric atoms is expected to affect the solar wind interaction with Mars, but its influences and significance have only been touched upon in previous work. Here several features associated with the charge exchange process between the solar wind protons and Martian neutral upper atmospheres are described. The analysis is based on an empirical proton model derived from Phobos 2 observations interacting with the Martian atomic (H) and molecular (H2) hydrogen, and oxygen (O) upper atmospheres representing solar minimum and solar maximum conditions. The region where the largest fraction of solar wind protons is lost by the charge exchange process is found to be a thin layer above the surface of Mars on the dayside resulting from charge exchange with the thermal oxygen. In general, the magnetosheath and magnetosphere (where the observed plasma takes on a different character in the Phobos 2 data) produce two distinguishable regions where the loss rate of solar wind protons is highest. Increasing solar activity increases the loss rate in the magnetosheath but decreases it in the magnetosphere. No significant increase of the absorption of the solar wind was found near the magnetopause suggesting that the decrease of the solar wind protons observed by Phobos 2 are not due to the charge exchange process. In addition to a reduction in the solar wind density, the charge exchange reaction results in energetic neutral atom (ENA) production. This paper considers some of the detailed properties expected for the ENA population at Mars. The ENA differential fluxes were found to be typically 106--107 cm-2 s-1 keV-1 in the energy range 0.01--1 keV. During solar minimum, the ENA production rate and ENA integral fluxes were found to be highest in the magnetosheath. At solar maximum the ENA production rate is highest in the magnetosphere, and ENA integral fluxes in the dayside magnetosphere appear to become comparable to the fluxes in the magnetosheath if the proton temperature in the magnetosphere is low. It is found that 1--3% of the original solar wind proton flux converts into ENAs before the bow shock. The ENAs produced upstream are undeflected and so may precipitate into the Martian upper atmosphere, depositing an energy flux of up to 3¿109 eV cm-2 s-1 derived from the solar wind. These results both suggest the possible benefits of observing ENA fluxes around Mars and suggest the necessary parameters for detector design. ¿ 1997 American Geophysical Union