Fine needle aspiration cytology was used widely to select thyroid nodules for surgery. The result could be highly reliable for most malignancies and for benign nodules. The purpose of this study was to determine the value of frozen biopsy by directly comparing the results of preoperative fine needle aspiration cytology to frozen biopsy examination. In our university hospital, 103 patients with thyroid nodule were operated during the years 1996 through 1997. A comparison of accuracy of the fine needle aspiration cytology with frozen biopsy was made for 85 patients who underwent both procedures. The 85 patients were separated into two groups. The group I included 65 patients, whose preoperative fine needle aspiration cytology results were reported as benign or malignant lesion definitely. The group II included 85 patients, the patients whose reported fine needle aspiration cytologic result was suspicious for malignancy were regarded .as malignant lesion and results for follicular neoplasm were regarded as benign lesion. The accuracy of both procedures was compared on two group also. The sensitivity, specificity, and accuracy of frozen biopsy were 76.7%, 93.5%, and 85.2% in group I and 75.6%, 95.5%, and 85.9% in group II, respectively, compared with 63.3%, 87.1%, and 75.4% in group I and 65.4%, 81.8%, and 72.9% in group II for fine needle aspiration cytology. On benign lesion, the accuracy was 95.5% in frozen biopsy and 68.2% in fine needle aspiration cytology, 75.6% in frozen biopsy and 53.7% in fine needle aspiration cytology on malignant lesion. We conclude that although fine needle aspiration cytology is the reliable diagnostic test in the evaluation of thyroid nodule, frozen biopsy should be performed intraoperatively, because fine needle aspiration cytology is insufficiently sensitive for reliable surgical planning.