This paper proposes a Hierarchical Time-Dependent Shortest Path (HTDSP) algorithm system that applies the hierarchical search strategy in dynamic traffic networks with discrete and deterministic travel costs. Four types of TDSP query experiments, namely, quasi-normal one-to-one queries, random one-to-one queries, one-to-all queries and all-to-one queries, were conducted on an urban street network of Tucson, Arizona. The numerical results suggest that the proposed algorithm provides near optimal solutions at low computational cost for both single and multiple sources TDSP queries. The flexibility and efficiency of the proposed HTDSP system implies its suitability for large-scale dynamic traffic assignment and online routing applications in which origin destination and departure time are required in a highly disaggregate manner.