The electron-impact excitation (EIE) cross-sections of 5s2S1/2 → 5p2P3/2 of Cd+ have been calculated by using the relativistic distorted-wave (RDW) method which we have developed recently. In order to discuss the electronic correlation effects, four different models are used to describe the target wave functions, namely Model A, 8S, 5SD and 6SD. Model A is a single configuration model, it includes 5s1/2, 5p1/2 and 5p3/2 only. The 8S is a valence-valence correlation model, it considers the impacts of 6s, 7s, 8s and 6p, 7p, 8p orbitals on wave function of 5s1/2, 5p3/2, respectively. The 5SD is a core-valence correlation model, it includes all the virtual single and double excitations from the 4s, 4p, 4d, 5s and 5p shells into the unoccupied 4f, 5p, 5d, and 5f shells. The 6SD is also a core-valence correlation model, it includes all the virtual single and double excitations from the 4s, 4p, 4d, 5s and 5p shells into the unoccupied 4f, 5p, 5d, 5f, 6s, 6p and 6d shells. The oscillator strength calculated by Model A is 0.72, which is in disagreement with the experiment 0.55 measured by Xu et al., while the results of 6SD 0.57 is in agreement with these of the experiment very well. For the EIE cross-sections, the core-valence correlation is very important. The results of Model A and 8S are larger than the experimental results obtained by Gomonai et al., while the results of 5SD and 6SD is obviously smaller than the results of Model A. In low energy range (2P3/2 → 5s2S1/2 emission lines after excitation are very small, while the linear polarization of Model A, 5SD and 6SD are in consistent with each other and also in good agreement with other theories. But for high energies, the theoretical results have big difference from the experimental results obtained by Goto et al.