The problem of searching the closest lattice point in large dimensional lattices finds many applications in single and/or multiple antenna communications. In this paper, we propose a Gaussian sampling based lattice decoding algorithm (GSLD). The algorithm iteratively updates each coordinate by sampling from a continuous Gaussian distribution and then quantizes the sampled value to the nearest alphabet point. The algorithm complexity per iteration is independent of the size of the alphabet, and hence is of high interest in higher order modulation schemes. We show that the algorithm is able to achieve near-optimal performance in polynomial complexity in different wireless communication system models.