Word sense disambiguation is essential for semantic analysis in many natural language-related applications, such as information retrieval, data mining, and machine translation. One of the effective models for word sense disambiguation is the word space model that represents context vectors and sense vectors in a word vector space. In this paper, we extend the word vector space model to reflect a more finegrained meaning in context vectors by incorporating embedded senses. Using a large Korean sense-tagged corpus, we build an embedded sense space with supervised learning and evaluate the effectiveness of the sense embedding for word sense disambiguation.