One of the main goals for designing database is minimize the data redundancy. Literature shows that there is several approach have been used to minimize the data redundancy, such as converting an Entity Relational Diagram (ERD) scheme according to steps of an algorithm for ER-to-relational mapping, and applying the normalization rules. These techniques have improved the database design quality, reduced the data redundancy, and omitted from a large proportion of repetition. However, there are still a big proportion of duplicated values especially for the huge database as some attribute are related to each other and may have the same data, such as the attributes of first_name, middle_name, and family_name. These attributes cannot be called as the multivalued attribute as these values belong to variety of entities or relations. Therefore, we propose a novel normalization forms for relational database design that match the related data attribute. This proposed approach called Matching Related Data Attribute Normal Form (MRDANF). A civil registration database system is used as a case study to validate the proposed approach. The results show that using our proposed approach has the positive impact on database quality and performance as the data redundancy will be reduced.