A random walk method for predicting percutaneous drug absorption pharmacokinetics was proposed. The profiles predicted by this method were compatible with those predicted by the analytical method. The random walk method is particularly useful for predicting complex processes such as repeated topical application of a drug. The amount of a drug released into skin from four therapeutic timolol patches was measured when the patches were serially applied for 2.5 h each on the same site of six healthy male volunteers. On the average, 33.2, 23.4, 15.1, and 16.5% of the applied dose was released into skin from the first, second, third, and fourth patches, respectively. This pattern was comparable with the predicted profiles (43.9, 30.2, 24.4, and 21.1%) of amounts of drug which were expected to be released from the first to fourth patches into skin, respectively. The estimation method for the normalized skin-capillary boundary clearance is also described and applied in examining the percutaneous absorption of timolol. The estimated value for this parameter was much greater than the diffusion parameter, indicating that the removal process of timolol by the local circulation is much faster than the diffusion process through skin.