The progress of grain boundary engineering (GBE) is overviewed and the challenges for further investigations emphasized. It points out that, the electron backscatter diffraction (EBSD) reconstruction of grain boundaries, which gives the information of connectivity interruption of general high angle boundaries (HABs), is more significant than purely pursuing high frequency of so-called special boundaries. The criterion for the optimization of grain boundary character distribution (GBCD) needs to be established. The energy spectrum and the degradation susceptibility of grain boundaries of various characters including HABs and low Σ（Σ≤29） coincidence site lattice (CSL) needs to be studied and ascertained. And finally, the newly proposed model of non-coherent Σ3 interactions for GBCD optimization are discussed.