ABSTRACT Glaucoma is the most common sources of the retinal disease that leads to permanent impaired vision worldwide. An automatic Optic Disc (OD) findings in retinal images utilized to diagnosis eye-related diseases like diabetic retinopathy. Numerous methods are offered to detect OD in low-resolution retinal images. This work presents an automatic glaucoma diagnosis using an image processing technique from the digital fundus image. In this work, a novel Particle Swarm Optimization (PSO) optimized KNN used for glaucoma disease classification. PSO is a naturally inspired optimization algorithm, utilized to find optimization parameters of KNN to improve classification accuracy. The proposed algorithm divided into three stages. Preprocessing stage includes noise removal, contrast enhancement using histogram equalization. For OD detection FCM has been used. Finally, PSO-KNN classifier used for categorizing healthy and non-healthy images of Optic Disc. The proposed technique has been coded in MATLAB and tested in the standard database of DRIVE and STARE fundus image. From the result observed that compared to other algorithms proposed approach improves accuracy considerably.