Abstract: To deal with the shortcomings of traditional swarm intelligence algorithm in solving the problem of SoC test multi-objective optimization, a combination of the improved chaotic Tent map and multi-objective genetic algorithm is presented. The model for optimization of both test time and test power is established. The model is solved with the proposed algorithm on the basis of the reasonable partition of TAM bus. The experimental results on the chosen circuits of ITC'02 benchmarks show that the proposed algorithm is effective and superior to the traditional ones.