In this paper, a recently proposed dimensionality reduction method called twin kernel embedding (TKE) is applied in 2-dimensional visualization of protein structure relationships. By matching the similarity measures of the input and the embedding spaces expressed by their respective kernels, TKE ensures that both local and global proximity information are preserved simultaneously. Experiments conducted on a subset of the structural classification of protein (SCOP) database confirmed the effectiveness of TKE in preserving the original relationships among protein structures in the lower dimensional embedding according to their similarities. This result is expected to benefit subsequent analyses of protein structures and their functions.