Lung cancer is one of the leading causes of mortality in both men and women throughout the world. That is why early identification and treatment of lung cancer patients bear a huge significance in the recovery procedure of such patients. A lot of time, pathologists use histopathological pictures of tissue biopsy from possibly diseased regions of the lungs to detect the probability and type of cancer. However, this procedure is both tedious and sometimes fallible too. Machine learning based solutions for medical image analysis can help a lot in this regard. The aim of this work is to provide a convolution neural network (CNN) model that can accurately recognize and categorize lung cancer types with superior accuracy which is very important for treatment. We propose a CNN model with 15000 images split into 3 categories: Training, validation, and testing. Three different types of lung tissues (Benign tissue, Adenocarcinoma, and squamous cell carcinoma) have been examined. 50 instances from every class were kept for testing procedure. The rest of the data was split as: about 80% and 20% for training and validation respectively. Eventually, our model obtained 98.15% training accuracy and 98.07% validation accuracy.