The matters of optimum forecasting atmospheric temperature using GPS radio occultation measurements are considered. The analysis of the available data regarding to the comparison of temperature measurements using radio occultation method and radiosondes was made. As a result it was concluded that the mean value of those results’ difference and also the mean quadratic deviation of these difference increases in common by increase of the forecasting time. In order to prevent surplus loading of telemetry channels and broadcasting inaccurate forecast values via them the optimization of general procedure of radio occultation temperature measurements are carried out using fine functions method. For optimization the concurrent parameters, changing on antiphase order are determined. It is found out that utilization of fine function method taking into account the applied optimization criterion and some limitation conditions make it possible to optimize the whole procedure of forecasting atmospheric temperature using the GPS radio occultation measurements.