Intervertebral disc spacers using bioactive ceramics have been used to treat degenerative spinal disease. Tooth-shaped spacers are commonly used to prevent migration, but there is a possibility of fracture when inserted or after insertion. Intervertebral disc spacers with either an isosceles triangle-shaped tooth (T1) or a right triangle-shaped tooth (T2) were used as a control group. The design factors for the experimental group were modified to prevent fractures induced by stress concentration, and the surfaces of the spacers were designed as either an isosceles triangle-shaped valley (V1) or a right triangle-shaped valley (V2). Linear analysis using finite element model (FEM) was performed, and Von Mises stress distribution was calculated by applying 1000 N of uniformly distributed load. Samples of the V2 design were made with bioactive glass-ceramics (BGS-7) and evaluated for compressive strength, fatigue degree, and impact strength. Von Mises stress was highest at the first tooth from the posterior side for the control group and at the center for the experimental group. Compared with the control group, the experimental group showed 18.4% and 82.5% reduction (V1 vs. T1 and V2 vs. T2, respectively) in the maximum stress at the bottom of the valleys. The FEM analysis revealed that the V2 design had the most even load distribution. The V2 samples with bioactive glass-ceramics were evaluated for compressive strength, and all six samples were not fractured up to 24 000 N. However, the average impact strength was 19.42 kN, suggesting that momentary force caused damage at a lower load than compression with a steady speed. The BGS-7 intervertebral disc spacer with V2 design was not fractured during the fatigue test at maximum pressure of 8000 N, R ≥10, 5 Hz, and 5 million cycles. These data confirm that the BGS-7 spacer with the V2 design may be clinically applicable. Collectively, the modified surface geometry of the experimental group significantly lowered Von Mises stress values at the bottom of the valleys, and thus the possibility of fracture by compressive load was greatly reduced. Also, impact during insertion was confirmed to cause fracture more easily, as the impact strength was lower than the compressive strength in the experimental group.