Objective   To investigate the influencing factors for catheter-associated infection (CRBSI), to construct a prediction model, and to build up treatment strategies of the injection in patients with maintenance hemodialysis (MHD).  Methods   A total of 204 MHD patients in our hospital were selected as the control subjects. They were divided into infected group (n=63) and non-infected group (n=141). Logistic regression was used to analyze the patients’ clinical data and to construct the model. A new intervention system was then established based on the model, clinical experience and literature reports. In addition, a total of 198 MHD patients were selected as the observation subjects to evaluate the effects of the new intervention system.  Results Diabetes history (OR=2.237, 95% CI: 1.067～5.679, P=0.003), catheter placement time (OR=3.647, 95% CI: 1.326～7.892, P=0.024), frequency of catheter care (OR=6.679, 95% CI: 2.534～16.476, P=0.003), aseptic training rate (OR=2.889, 95% CI: 1.146～6.654, P=0.030) and blood uric acid (OR=2.937, 95% CI: 1.627～8.883, P=0.014) were the independent influencing factors for CRBSI in MHD patients. The model equation was Y=-20.673+1.116×blood uric acid+1.012×aseptic training rate+2.017×catheter care frequency+1.912×catheterization time+1.079×diabetes history. Hosmer and Lemeshow's goodness of fit test equation was used (χ2=10.678, P=0.236). Receiver operating characteristic curve showed that the area under the curve was 0.813, the sensitivity was 89.47%, and the specificity was 63.93%. The incidence of CRBSI and adverse events were lower in the observation group than those in the control group (χ2=3.606 and 32.780; P <0.001).  Conclusion  The prediction model has a better discrimination validity, and the evidence-based systematic nursing intervention system has a better effect on preventing and controlling CRBSI in MHD patients.