The symmetric implosion of the pellet is very important in an inertial confinement fusion (ICF) system. The evaluation of symmetric distribution on the ICF pellet is time and memory consuming, especially when the elements of cylindrical hohlraums are iteratively subdivided. A novel compressive sensing (CS) based approach is presented in this paper to accurately valuate the distribution symmetry with much less computation. The core idea is that, the thermal radiation distribution is transformed in the spherical harmonic field, and the sparsity of spherical harmonics is accurately evaluated through CS approach. Finally, experiments are demonstrated to show higher efficiency of the proposed approach.