Unrelenting demand of mobile data can be met by intensive spectrum and spatial reuse which can be achieved by femtocells and multi-stream MIMO transmission. However, random deployment of femtocells may cause severe co-channel interference (CCI). In this paper we propose an algorithm, where femto base stations form a cluster and cooperatively generate precoding matrix for interference mitigation to nearby macro user (MUE) and other femto users (FUEs). We propose a modified version of conventional block diagonalization (cBD) linear precoding, where some antennas in femto cluster are de-activated for interference alignment in the null space of MUE. The cBD involves two singular-value- decomposition (SVD) operations which introduces high computational complexity. Therefore we propose a low complexity precoding algorithm which involves generalized zero forcing channel inversion (ZF-CI), QR decomposition and lattice reduction (LR) transformation. Simulation and analytical results show the superior performance of our proposed scheme in terms of sum-rate and computational complexity.