In this paper, a demonstration of DNA computing for k-shortest paths of a weighted graph is realized by biochemical experiments in such a way that every path is encoded by oligonucleotides and the length of the path is directly proportional to the length of oligonucleotides. For initial pool generation, parallel overlap assembly is employed for efficient generation of all candidate answers. During the computation, after the initial pool of solution is subjected to amplification, which is polymerase chain reaction, k-shortest paths could be separated by gel electrophoresis and the selection can be made at last