A 5-point harness type of child seat approved by ECE R44 was studied in this paper. This child seat can be installed either by the ISOFIX anchorages (the ISOFIX CRS) or by a vehicle 3-point seatbelt (the universal CRS). Hybrid III 3-year-old (3YO) dummy restrained by the ISOFIX CRS was used in a vehicle buck sled test. TNO P-series 3-year-old (P3) dummy restrained by the universal CRS was used in a frontal 50kph full rigid barrier (FRB) test. Multi-body models of the two CRSs were developed by using Madymo software and correlated with the two tests. The goals of this study are to compare the performance of the ISOFIX CRS and the universal CRS, to understand the discrepancy of the kinematics of the Hybrid III 3YO and P3 in the multi-body model simulations, and to find the influence factor of a pulse on the dummy responses. Five different crash pulses were used in the Madymo simulations for both the ISOFIX CRS and the universal CRS. The head forward excursion, and the head, chest, and pelvis accelerations of the Hybrid III 3YO restrained by the ISOFIX CRS were lower than those restrained by the universal CRS when being applied by the same crash pulse. The head forward excursion of P3 was about 10 mm greater than that of the Hybrid III 3YO in the simulations of the five crash pulses. Trapzoidal wave approximation (TWA) is a useful method to approximate the crash pulse. TWA pulse has the same dynamic crush and the time when obtaining the maximum dynamic crush with the crash pulse. TWA pulses were calculated based on the five crash pulses. The dummy responses were similar in the simulations of the crash pulse and the calculated TWA pulse. It was found that the dummy accelerations decreased when the rising slope of the TWA pulse decreased.