A CT reconstruction method based on fractal and compressed sensing, including steps: (1) known CT projection data b and corresponding projection angle θ, initialization parameters β, λ, λn, CT image u=0; (2) according to the projection Angle θ, calculate the corresponding projection matrix A; (3) Use formula (5) to calculate CT image uART; (4) Perform fractal coding on uART to get Φ(u), use formula (12) to get α; (5) Perform decoding processing on Φ(u) to get ΦT(u); 6) Solve ΦT(u)Φ(u), ΦT(u)α, ATA, ATb respectively; (7) Use conjugate gradient method to solve equation (15) , get the CT image 8) Check whether the condition for the end of the iteration is met, go to step (9), otherwise step (3); (9) ends, the output CT image method uses fractal as a sparse transformation to obtain a more sparse coefficient representation, Iterative continuous iterative algorithm is used to solve the problem, and high-quality CT images can be reconstructed with less projections.