In this paper, tensile behavior of joint filling has been investigated under experimental test and numerical simulation (particle flow code). Two concrete slabs containing semi cylinder hole were prepared. These slabs were attached to each other by glue and one cubic specimen with dimension of 19 cmx15 cmx6 cm was prepared. This sample placed in the universal testing machine where the direct tensile stress can be applied to this specimen by implementing a special type of load transferring device which converts the applied compressive load to that of the tensile during the test. In the present work, two different joint filling thickness i.e., 3 mm and 6 mm were prepared and tested in the laboratory to measure their direct tensile strengths. Concurrent with experimental test, numerical simulation was performed to investigate the effect of hole diameter, length of edge notch, filling thickness and filling length on the tensile behavior of joint filling. Model dimension was 19 cmx15 cm. hole diameter was change in four different values of 2.5 cm, 5 cm, 7.5 cm and 10 cm. glue lengths were different based on the hole diameter, i.e., 12.5 cm for hole diameter of 2.5 cm, 10 cm for hole diameter of 5 cm, 7.5 cm for hole diameter of 7.5 cm and 5 cm for hole diameter of 10 cm. length of edge notch were changed in three different value i.e., 10%, 30% and 50% of glue length. Filling thickness were changed in three different value of 3 mm, 6 mm and 9 mm. Tensile strengths of glue and concrete were 2.37 MPa and 6.4 MPa, respectively. The load was applied at a constant rate of 1 kg/s. Results shows that hole diameter, length of edge notch, filling thickness and filling length have important effect on the tensile behavior of joint filling. In fixed glue thinks and fixed joint length, the tensile strength was decreased by increasing the hole diameter. Comparing the results showed that the strength, failure mechanism and fracture patterns obtained numerically and experimentally were similar for both cases.