There are lots of Super resolution methods developed recently. Each has its own pros and cons and behavior. The neighbor-embedding (NE) algorithm for single-image super-resolution reconstruction is one of them which assume that the feature spaces of low-resolution and high-resolution patches are locally isometric. Even, this is not true for SR because of one to many mappings between Low Resolution and High Resolution patches. To minimize the problem for NE-based SR reconstruction, an advanced Neighbor Embedding based method for Super resolution used in which combine learning technique used to train two projection matrices simultaneously and to map the original Low Resolution and High Resolution feature spaces onto a unified feature subspace. Reconstruction weights of the k-Nearest neighbour of Low Resolution image patches is found by performing operation on those Low Resolution patches in unified feature space. To handle a large number of samples, combine learning use a coupled constraint by linking the LR–HR counterparts together with the k-nearest grouping patch pairs. The Advanced NE algorithm gives better resolution and outperforms NE method for image super resolution.