Dynamic loads, including earthquake, blasting, and vibration loads, induce cyclic shear loads along the joints in rock masses; hence, the risk of failure increases on the joints due to changing shear resistance. On the other hand, joints are under different boundary conditions: constant normal load (CNL) and constant normal stiffness (CNS). Normal stiffness increases on the joints with increasing depth, and it can affect shear resistance. For an accurate assessment of joint shear resistance under varying normal stiffness and number of cycles, advanced laboratory shear apparatus is essential for the shear test. Conventional direct shear apparatuses have limitations such as boundary conditions, working under monotonic (static) shear loads only, or cyclic shear loads with no change of frequency and amplitude of shear loads. Therefore, a new large-scale servo-controlled direct shear testing machine was developed to conduct cyclic shear test (as well as monotonic shear test) under CNL and CNS boundary conditions with varying normal stiffness at different frequencies and amplitudes of shear loads. In the present study, the cyclic shear tests were conducted on nonplanar joints under varying normal stiffness. Moreover, the effects of different frequencies and amplitudes of shear loads were investigated. The test results indicate that peak shear stress increases with increasing normal stiffness at the first cycle, but the influence of normal stiffness decreases with an increase in the number of shear cycles. The frequency of shear load influences peak shear stress, i.e., peak shear stress increases with increasing frequency. The number of cycles does not affect peak shear stress on the joints at low shear amplitude, but peak shear stress decreases with higher amplitude.