Background: Central venous catheters (CVCs) have been essential devices for the treatment of children with hematological and oncological disorders. Only few studies investigated the complications and selections of different types of CVCs in these pediatric patients. This study aimed to compare risk factors for unplanned removal of two commonly used CVCs, i.e., peripherally inserted central catheters (PICCs) and tunneled CVCs, and propose better device selection for the patient. Procedure: This retrospective, single center cohort analysis was conducted on pediatric patients with hematological and oncological disorders inserted with either a PICC or a tunneled CVC. Results: Between January 1, 2013, and December 31, 2015, 89 patients inserted with tunneled CVCs (total 21,395 catheter-days) and 84 with PICCs (total 9,177 catheter-days) were followed up until the catheter removal. The median duration of catheterization was 88 days in PICCs and 186 days in tunneled CVCs (p = 1.24×10-9). PICCs at the 3-month cumulative incidence of catheter occlusion (5.2% vs. 0%, p = 4.08×10-3) and total unplanned removal (29.0% vs 7.0%, p = 0.0316) were significantly higher, whereas no significant difference was observed in the cumulative incidence of central line-associated bloodstream infection (11.8% vs. 2.3%, p = 0.664). Multivariable analysis identified younger age (<2 years) (subdistribution hazard ratio [SHR], 2.29; 95% confidence interval [CI], 1.27–4.14) and PICCs (SHR, 2.73; 95% CI, 1.48–5.02) were independent risk factors for unplanned removal. Conclusion: Our results suggest that tunnel CVCs would be a preferred device for children with hematological and oncological disorders requiring long-term, intensive treatment.