A method of generating information on relationships between characters of a content includes dividing a text extracted from the content into one or more predetermined units, determining one or more dominant relationships between characters of the content by comparing the divided units with relationship keyword information in which keywords contained in categories are defined, wherein the categories represent one or more relationships between the characters, and generating information on the relationships between the characters in accordance with the determined dominant relationships.The dominant relationships are determined by matching the divided units of text to the categories with reference to the relationship keyword information, counting the number of divided units of text corresponding to each of the categories, and determining the relationship represented by the category measured by the highest number of divided units of text.