To determine the permeabilities from a vertical interference test, earlier investigators proposed correlations or type curves based on point source solutions or partially penetrated well solutions. However, it is impossible for these correlations or type curves to cover all possible scenarios in the real world. In more recent years, people use regression analysis to simultaneously match the pressure responses at the source interval and the observation point. With regression analysis, we face the problem of non-unique solutions. Sometimes, estimated permeability could be outside a reasonable range when analyzing noisy data from real tests. In this paper, a new technique is presented to estimate horizontal and vertical permeabilities from the time-delayed response in a wireline vertical interference test. In a vertical test, a pressure drawdown test period is followed by a pressure buildup test period. Because of the delay in response, the pressure at the observation probe continues to drop for some time while the pressure at the source interval is being built up. The maximum pressure drop at the observation probe with regard to the maximum pressure drop in the source interval is time-delayed. Using this time delay and the maximum pressure drop at the observation probe, vertical and horizontal permeabilities can be estimated. A novel numerical scheme is used. The new technique is compared with the previous methods, and it shows its superiority in accuracy. In addition, it can be used in different test configurations. Application issues in real testing conditions are discussed. Finally, two field tests are analyzed using this technique, while an earlier effort to analyze the tests using a conventional method was not successful due to poor data quality.