The use of three-dimensional preoperative planning and bone tumor resection guided by navigation has increased in the last ten years. However, no study to date, as far as we know, has directly provided evidence of accuracy of this method. The objective of this study was to describe a method capable of determining the accuracy of osteotomies performed for tumor resection planned and guided by navigation. We hypothesize that matching the 3D reconstructed surgical specimen is an acceptable method to determine the accuracy of virtual planning and navigation. A total of seven patients and 14 osteotomies were evaluated. After surgery, all surgical specimens were 3D reconstructed from CT images. The mean of quantitative comparisons between osteotomies planned and osteotomies obtained through the resected specimen was in a global mean of 1.56 millimeters (SD: 2.91) for all the cases. Based on our observations, a three-dimensional model obtained from the tumor surgical specimen is a useful tool to determine accuracy of 3D planning and surgical navigation.