Chronic Kidney Disease nowadays is a very harmful disease leading to a large number of deaths around the world. Persons suffering from this disease are increasing day by day. Factors like age, geographic region, gender, economic conditions and lack of water consumption can result in the occurrence of this disease. As this disease has very expensive treatment if not treated properly at an early stage, it can lead to dialysis and kidney transplant too. If disease diagnosis is done manually, it becomes difficult for the doctors to make an exact prediction about the disease based upon patients’ medical history, diagnostic tests and procedures that define the symptoms. Due to technological advancement, accurate disease diagnosis and its treatment can be done easily. The prediction of chronic disease can be done at early stages so as to overcome and prevent the disease. For this a model that is based upon a patient's requirement is highly recommended. Although there are many other diagnostic techniques, Machine Learning, is given priority because of its great accuracy and is thought to be the most trustworthy diagnosis method since it minimizes human error. Various Machine Learning algorithms are available that can be applied on the disease dataset and find which one is accurate based upon different attributes and parameters. This paper discussed all the different tools, techniques, algorithms and hybrid approaches to detect Chronic Kidney Disease and compared them in terms of their accuracy.