Restorative procedures can lead to weakening tooth due to reduction and alteraton of tooth structure. It is essential to prevent fractures to conserve tooth. Among the several parameters in cavity designs, cavity isthmus and depth are very important. In this study, MO amalgam cavity was prepared on maxillary first premolar. Three dimensional. finite element models were made by serial photographic method and cavity depth(1.7mm, 2.4mm) and isthmus (11 4, 1/3, 1/2 of intercuspal distance) were varied. linear, eight and six-nodal, isoparametric brick elements were used for the three dimensional finite element model. The periodontal ligament and alveolar bone surrounding the tooth were excluded in these models. Three types model(B, G and R model) were developed. B model was assumed perfect bonding between the restoration and cavity wall. Both compressive and tensile forces were distributed directly to the adjacent regions. G model(Gap Distance: 0.000001mm) was assumed the possibility of play at the interface simulated the lack of real bonding between the amalgam and cavity wall (enamel and dentin). When compression occurred along the interface, the forces were transferred to the adjacent regions. However, tensile forces perpendicular to the interface were excluded. R model was assumed non-connection between the restoration and cavity wall. No force was transferred to the adjacent regions. A load of 500N was applied vertically at the first node from the lingual slope of the buccal cusp tip. This study analysed the displacement, von Mises stress, 1 and 2 direction normal stress and strain with FEM software ABAQUS Version 5.2 and hardware IRIS 4D/310 VGX Work-station. The results were as follows: 1. G model showed stress and strain patterns between Band R model. 2. B model and G model showed the bending phenomenon in the displacement. 3. R model showed the greatest amount of the displacement of the buccal cusp followed by G and B model in descending order. G model showed the greatest amount of the displacement of the lingual cusp followed by B and R model in descending order. 4. B model showed no change of the displacement as increasing depth and width of the cavity. G and R model showed greater displacement of the buccal cusp as increasing depth and width of the cavity, but no change in the displacement of the lingual cusp. 5. As increasing of the width of the cavity, stress and strain were not changed in B model. Stress and strain were increased on the distal marginal ridge and buccopulpal line angle in G and R model. The possibility of the tooth fracture was increased. 6. As increasing of the depth of the cavity, stress and strain were not changed in B and G model. Stress and strain were increased on the distal marginal ridge and buccopulpal line angle in R model. The possibility of the tooth fracture was increased.