In order to a nalyze t he steady -state perf ormance of the P arz en probabil ity dens ity functi on ( PPDF) algorithm based o n th e Pa rze n win dow metho d, the numeri cal solu tion method wa s use d to input fou r ki nds of modulati on sig nal s re-specti vely, and the va lue of th e comp ensation fa ctor of the PPD F a lgorithm u nder differe nt kernel p arameters w as stud-ied. In ad dition , the closed analyti cal ex pressio ns of the steady -sta te mean sq uar e error (MS E) o f the PPDF algo rithm unde r the con diti ons of real -valued an d complex- valued data wer e obtained by usi ng the sep arat ion hypoth esis pri nc iple. The simulation results show that the theoretical values of real and com plex  stea dy MSE of  PPDF  alg orit hm are in  good agreement with the simulation values in 
the noiseless environment, which verifies the accuracy of the theoretical analysis.