A multiexponential allometry (MA) method was developed to predict human drug clearance from preclinical data. Separate data sets containing clearances from human and preclinical species were chosen for the study. Human clearance was estimated using the MA technique according to the equation: CL = aBWb + cBWd, where CL is clearance in milliliters/minute, and a, b, c, and d are constants derived from preclinical pharmacokinetic data. Simple allometry (SA) gave the poorest prediction using any data set, and the percentage outliers remained larger than MA or monkey liver blood flow within 1.5-, 2-, and 3-fold error. Analysis of compounds common to both data sets suggested that MA could accurately predict human clearances within approximately 10% of 3-fold error. The analysis also showed that monkey is an important species for scaling, and MA is a better predictor of human clearance when the slope of SA is >0.7.