Distributed computing environment composed of interconnected machines with varied or same computational capabilities is well suited to meet the computational demands of diverse groups of tasks. The most popular model characterizing tasks' dependence is to utilize DAG(directed acyclic graph). In paper [12], a novel model called TTIG that is more realistic and universal than DAG and its corresponding algorithm called MATE for static mapping of parallel application are presented. We extend TTIG model, and propose a new static scheduling algorithm called GBHA (Group-based Hybrid Algorithm) and two versions (GBHA1 for homogeneous systems and GBHA2 for heterogeneous systems). In this work, our algorithms are compared with MATE and some well-known scheduling algorithms for multiprocessor systems based on DAG. The simulation experiment results show that our algorithms outperform MATE significantly in both homogeneous and heterogeneous systems and can be comparable to efficient scheduling algorithms based on DAG in multiprocessor systems but with much lower complexity.