A semi-empirical model, Johnson's smog production model (SFM), which relates precursor emissions to ozone levels and estimates the relative effectiveness of volatile organic compounds (VOC) and NO X emission controls, has been evaluated and a modified version of SPM has been introduced. Both versions have been applied to routine data from 1989-1991 in five areas in the United States. In particular, extent parameters, which reveal the relative merit of VOC and NO X controls in reducing high ozone levels, have been calculated. Preliminary applications of SPM reveal interesting features with respect to VOC vs. NO X controls in reducing high ozone levels. For hourly data with ozone >0.08 ppm, distributions of extent parameters resulting from the modified SPM show the effectiveness of VOC controls at more monitoring sites than those from Johnson's SPM; however, relative features between the two versions are similar. On the other hand, for hourly data with ozone >0.12 ppm, the two SPM versions show very similar relative effectiveness of VOC and NO X controls with chosen values of model parameters. To improve the credibility of SPM, the range of validity of relationships between maximum smog produced or maximum ozone and NO X concentrations must be determined, and the parameters in these relationships must be better determined for typical VOC mixtures. Another essential parameter, which determines the fractional loss of NO y (NO and its oxidation products) from the gas phase must be better determined.