The current specifications for selecting crack sealants correlate poorly with actual field performance. To address this issue and assist in predicting the low-temperature properties of hot-poured bituminous crack sealants, a modified direct tensile tester method has been developed. Sample geometry is modified to accommodate testing sealants. A sensitivity analysis considering various loading rates, sample lengths, and cross-section areas was conducted to define both optimized specimen geometry and testing protocol. Two types of sealants, having a wide range of rheological behaviors (one polymer-modified and one having crumb rubber), were tested at low temperature. Results showed that the rich polymer-modified sealant has a high resistance to failure compared with the sealant with crumb rubber-modified. Each sealant was tested at the lowest corresponding expected service temperature. A performance parameter, strain energy density, was proposed to differentiate crack sealant material in the laboratory.