The study objective is evaluation of a benchmark approach for predicting mainstream smoke constituent machine-yields for conventional cigarette brands from worldwide markets. Results for ISO smoke yields support the validity of benchmarking when brands, for which yields are to be predicted, have design characteristics within boundaries established by the exploratory brands. Yields of ISO-method mainstream smoke constituents were generally well described by weighted least squares regression relationships with ISO tar (R2>0.80 and coefficient p values <0.05). The impact of the varied chemical composition of cigarette tobaccos from different regions on smoke constituent yields was recognized. Mainstream smoke nitrogen oxides and tobacco-specific nitrosamine (TSNA) yield prediction relationships improved by including tobacco nitrate or TSNA concentration factors in respective independent parameters. For carbon-filter brands, inclusion of a carbon factor improved the predicting relationships for several vapor-phase constituents. Relationships were validated with a subset of additional validation brands. Greater than 90% of the validation brands' smoke chemistry yields were within the 95% prediction intervals. Average differences between measured and predicted yields were generally within the range of one to two measurement standard deviations. The estimation methods proposed relate to machine-smoking conditions and are not intended to reflect the actual exposure of any given consumer to smoke constituents.