Thyroid lesions are very common problem in Bangladesh. To categorize the type of lesion, FNAC is a reliable, minimally invasive and cost-effective diagnostic modality. In routine cytology, precise categorization into benign and malignant is highly subjective. So, additional methods are needed to improve the accuracy of cytological diagnosis. Computerized semi-automatic nuclear morphometry is an objective and reproducible tool for quantitative evaluation of nuclear features which may help to differentiate between benign and malignant thyroid lesions in cytology. Objective of the study was to apply morphometry with regard to nuclear size parameters, shape parameters and nuclear chromasia with derivation of suitable cutoff values to differentiate benign and malignant lesions of thyroid gland. This cross-sectional study was done in total 55 cases at the Department of Pathology, Bangabandhu Sheikh Mujib Medical University, Bangladesh from March 2018 to February 2020. Morphometry was done on images containing diagnostic cells captured from cytological smears with open source ImageJ morphometric software. In this study, all the nuclear size parameters and one nuclear shape parameter was found significant in differentiating benign and malignant lesions. No significant difference was observed in analysis of chromasia parameter. The cut-off values derived between all benign and malignant lesions for mean value of nuclear area, maximum feret diameter, perimeter and aspect ratio were 40.20μm², 8.46μm, 6.68μm, 28.28μm and 1.35 respectively with sensitivity of 86%, 95%, 81%, 90% and specificity of 97%, 93%, 97% and 97% respectively. The Nuclear morphometry may be used as an adjunctive tool to routine thyroid cytology.