Current crack sealant specifications utilize a simple empirical bond test [ASTM D5329, "Standard Test Methods for Sealants and Fillers, Hot-Applied, for Joints and Cracks in Asphaltic and Portland Cement Concrete Pavements"] to measure a sealant's ability to adhere to concrete briquettes. However, no correlation exists between the results of this standard test and a sealant's field performance. To bridge the gap between sealants' fundamental properties and field performance, performance-based guidelines for the selection of hot-poured crack sealants have been developed. This paper proposes a new test procedure to help assess the capability of sealants to adhere to the sides of cracks in pavement. The proposed crack sealant adhesion direct tensile (CSADT) test procedure calls for the use of a direct tensile tester with modifications to the end pieces and the specimen holder. Sealant is confined between two aluminum pieces, which will move apart at 0.05 mm/s until failure occurs. Failure load and the displacement at failure are recorded and then used to assess the sealant's adhesion capability. A crack created on one side of the assembly during specimen preparation determines the initial failure location and ensures that the failure occurs at the interface and not within the sealant. The specimen preparation is designed to simulate sealant installation, and experimental conditions have been developed to represent crack openings during pavement contraction in cold weather. The repeatability of the suggested procedure was acceptable with an average coefficient of variation of 10.9 %. No significant difference existed between two operators or between two test setups.