Models of soil temperature in the literature address several broad groups of variables: first, those that are site-specific such as latitude and soil series, second, those that directly or indirectly measure insolation by proxies such as cloud cover, cover crops, and the sun's relative position to the location, and third, daily weather variables such as air temperature and precipitation. Although each of the models reviewed was effective in predicting daily soil maximum and minimum temperatures, a simpler model was developed that shows equal or better predictability to existing models using easily obtainable data, such as minimum (Amin) and maximum (Amax) air temperature and day of year. Five locations in Arkansas with about 15 years of observations form the basis of the data set. Two general models were developed to predict daily soil minimum and maximum temperatures at 5 and 10 cm. All parameters were significant in the general model (for all locations), while 10 of the 12 terms in the site-specific model were significant. Values ofR2for the models ranged from 0.911 to 0.951, indicating that the parameters used in each model consistently accounted for at least 91% of the observed variation in soil temperature. Only minor differences existed in the values ofR2when comparing the site-specific with the general model, suggesting that including variables for location does not sufficiently increase the prediction capability of the model enough to warrant their inclusion. The general model provides adequate predictability with a greater degree of efficiency. Hence, the general model may be used with appropriate variables (Amax, Amin, and day of year) to predict soil minimum or maximum temperatures at either the 5− or 10-cm depth. Two applied examples are presented that use the kind of information generated by these models.