In recent years, expression of color has gained much more importance in the digital environment involving the computer and the internet, and with the advance of the economy our surroundings have developed into harmonious colors. However, research on color harmony has been treated as an area of experts involved in art and design. Accordingly, the public frequently experience trouble in color selection without the help of experts because they lack knowledge regarding color harmony. Previous researches are able to propose harmonious colors but do not use systematic computer modeling. They rather confuse users in final decision making as they do not reflect individual preferences in color recommendation. Thus, we propose the method of extracting binary color harmony based on color harmony theory to recommend a harmonious binary color to a given color, and also the method of recommending binary color harmony based on user preferences to deduce what kind of harmony a user prefers. To do this, we first, for the method of extracting binary color harmony based on color harmony theory, we store color information of NCS into the database and extracted harmonious colors of a given color by proposing a way to express Ostwald Color Harmony theory in a relational algebra for query processing. Second, for recommending binary color harmony based on user preferences, we researched to find out what kind of color harmony a user prefers to each color. Based on the results, we applied collaborative filtering method to the color harmony recommendation method and proposed a way to recommend color harmony in the order of priority which reflects user preference according to user preferences.