The task for chromosome karyotyping and diagnosis is requiring repetitive, time consuming job and high cost even it is done by well-experienced cytogenetists. Therefore an web based chromosome karyotyping instruction system has been established to be able to analyze chromosomes and obtain necessary advises from the database instead of human experts and the database is including 2 divisions with database and agent.For the first of all, database model was constructed with relational database consisting of Patient_DB, image_DB, Disease_DB and Manage_DB. As the second procedure, knowledge base by IF THEN production rule was implemented to a knowledge domain with normal and abnormal chromosomes. For the last, independent agent with the inference by knowledge base could enter the inference data into the database.Experimental data were composed of normal chromosomes of 2,736 patients' cases and abnormal chromosomes of 259 patients' cases that have been obtained from GTG-banding metaphase peripheral blood and amniotic fluid samples.The completed system provides variously morphological information by analysis of normal or abnormal chromosomes and it also makes users enable to control and search the information in a short period with learning of high amount of knowledge.