The invention provides a source lossy compression coding method realized through combining a polar code with arithmetic coding and belongs to the communication field. According to the method, distortion performance of a system is guaranteed through utilizing polarization Z (an expression described as the specifications) of the polar code, and better compression performance is realized through selecting arithmetic coding; under limited length condition, arithmetic coding can approximately reach to lossless compression capacity, the theoretical performance of the method is different from asymptotic performance of polarization Z (an expression described as the specifications) of the polar code, arithmetic coding can further reach excellent performance at the limited length. The structure combines the polar code with arithmetic coding, polarization and coding are simultaneously carried out, probability utilized based on the polar code is still utilized during arithmetic coding compressionfor calculation; the polar code is combined with arithmetic coding to realize source lossy compression coding to realize better performance, complexity of arithmetic coding is linear relative to codelength, so total complexity is almost not influenced by utilizing arithmetic coding.