Computed tomography (CT) plays a crucial role in medical imaging, such as in diagnosis and therapy. The radiation from CT will result in excess radiation exposure to patients, hence the reduction of radiation from CT is essential. The proposed research focused on combining expectation maximization (EM) and nonlocal total variation (NLTV) for CT image reconstruction. This method reconstructs better quality image using less number of projection views corrupted by Poisson noise, thus reducing the overall radiation exposure. NLTV regularization reduces the effect of noise and improve the recovered image quality. The numerical results show the efficiency of the proposed work in comparison with other existing reconstruction techniques.