Pseudo dinucleotide composition (PseDNC) and Z curve are two widely used feature extracted methods from DNA sequences,and both of them show excellent performance in the classification issues for nucleotide sequences.Phase-specific PseDNC (psPseDNC) was proposed by adopting the advantage of the two methods in this work.We used recombination spots prediction as a case to illustrate the capability of psPseDNC and also PseDNC fused with Z curve theory based on a novel machine learning method named large margin distribution machine (LDM).