Near set theory presents a fundamental basis for observation, comparison and classification of perceptual granules. Soft set theory is proposed as a general framework to model vagueness. The purpose of this paper is to combine these two theories in what are known near soft sets in defining near soft topology based on a nearness approximation space.