Cervical cancer is the second most common cancer in women after breast cancer, causing the death of one woman every two minutes in the world. The most important risk factor originating from the cervix is infection with the papilloma virus (HPV). Cervical cancer screening programs are extremely important to reduce the incidence and death rates of this cancer. The primary goal of screening for cervical cancer is the accurate detection and timely treatment of intraepithelial precursor lesions of the cervix, in order to prevent cervical cancer. With the PAP smear test, cells in the cancerous stage are detected in the endocervical canal, and cancer development can be prevented before the cells turn into cancer with cancer treatment at this stage. The PAP test, which is used in early diagnosis, is an easyto-apply, low-cost, harmless, high-sensitivity test that also reduces the burden of treatment. Recent developments in the field of artificial intelligence have achieved serious success in the diagnosis of cervical cancer. In this study, a transfer learning-based cervical cancer detection method and an application developed to easily perform these procedures are presented. Cancerous and non-cancerous cervical cells were classified using pre-trained networks. Five popular pre-trained networks, namely Xception, VGG-16, DenseNet, InceptionV3, and InceptionResNetV2, were used for the problem and the obtained performance results were compared. In addition, an application has been developed so that experts working in this field can easily make such classifications. With this application, users can create their own models by conducting a new training, use the model created in this study, and quickly test which class the newly obtained images belong to. As a result of the study, DenseNet network obtained the highest accuracy with 94.72% accuracy. Experimental results show that the proposed approach can provide an inexpensive and rapid decision support system for cervical cancer detection that anyone can apply.