Abstract: In solving high-dimensional optimization problems, the conventional binary nding of genetic algorithm may occupy huge computer memories, need a large amount of decoding work, and evenaffect the implementation of genetic algorithm since the model parameter amount is considerably large.In light of these problmes, this paper presents a practical, efficient implidt binary coding scheme. Whenapplial to high-dimensional inverse problems, this coding method can not only reduce the memories occupied by parameters to the minimum, but also decrease the decoding work amount by a big margin.