Cervical cancer is the second most common cancer in women globally, it is the leading cause of female death, next to breast cancer. Sexually transmitted virus, known as Human papillomavirus, causes this cancer. This preventable diseases cause female death because of lack of cervical screening in health institutes. Cervical screening used to detect the precancerous lesion before developing cancer cells. Pap smear is one of cervical cancer screening techniques that uses microscope to visualize the cervix lesion or cervix cancer. However, visual inspection suffers from false positive or false negative results due to human errors. Deep learning algorithm currently used for the detection and diagnosis of cervical cancer. The ultimate objective of this thesis is to detect precancerous lesion before developing cancer cells using multi-class classification, on local Pap smear image data. In this thesis, 1224 Pap smear image collected from local health institute and annotated by pathologist. Preprocessing mainly focused on image denoising using bilateral filter to remove Poisson noise. Convolutional Neural Network (CNN) and Pre-trained VGG19 algorithms were developed using train, validation and test data split. From Classical machine learning, Support Vector Machine (SVM) and Random Forest (RF) algorithms were developed. The performance metrics of accuracy, precision, specificity and sensitivity registered as: 99%, 98.6%, 100%, and 98.5%, respectively, for CNN; 100%, 100%, 100%, and 100%, respectively, for VGG19; 96%, 94%, 93% and 100%, respectively, for SVM; 100%, 100%, 100%, and 100%, respectively, for RF. Pre-trained VGG19 and Random Forest models were outperformed. Hence, these developed algorithms can improve the diagnosis and detection of cervical lesion and cancer cells from Pap smear images, which are safe, simple, available and routine screening methods in cervical diagnosis to improve the quality of life of patients. Keyword: Bilateral, Carcinoma, Classical Machine Learning, Deep Learning, Denoising, Optimizer, Overfitting, Papilloma, Pap smear, Precancerous lesion, Transfer learning.