In agriculture, paddy is one of the major food grains in the Asian countries like China, India, Indonesia, Bangladesh and many more. But diseases on paddy like a Rice Blast, Bacterial Blight, and Brown Spot etc are causing huge damage on yielding of paddy. If the diseases are not detected at an early stage, then there will be a decrease in the production of paddy. The main objective of this paper is to develop and implement an algorithm for diagnosing paddy diseases at the early stage, which are Blast Disease(BD), Brown –Spot Disease(BSD), and Narrow Brown–Spot Disease(NBSD). This paper provides a method for detection of paddy diseases using image processing techniques, to recognize the diseases in paddy fields from images, based on color, texture and shape of diseased paddy automatically, and give the suitable solutions to the farmers, so that paddy diseases can be prevented at early stage and hence high yielding of paddy can be achieved.