Lupus is one of the most serious disease found mostly in the women. Most of the patients of Lupus are infected with Lupus nephritis. Many studies have been carried out so far by the medical experts for proper diagnosis and treatment manually. Some studies are intended towards diagonally the disease at an early stage by using gene dataset and most of the diagnosis is based on symptoms of the disease. But it is very difficult for the medical experts to diagnose the disease at an early stage to treat proper treatment and management could be provided to the patient. This paper is an approach towards the formation of a new dataset by augmenting the already existing gene dataset with the symptom dataset that has been collected from various medical institutions. The new formed dataset is presented to a MLP which has been developed here as this is capable of automating the staging of Lupus. This method minimizes the time and effort to diagnose the disease with great accuracy, when compared with the already existing ML methods viz SVM, CNN. It has been found that recommendation wise the method developed here is better.