In this study, the effect of a flow behavior is taken into account to determine which type of a pulsating heat pipe (PHP) performs better between a closed-loop pulsating heat pipe (CLPHP) and a closed-end pulsating heat pipe (CEPHP). The thermal performances of a CLPHP and a CEPHP are experimentally evaluated and compared. For this, MEMS techniques are used to fabricate silicon-based PHPs with ten turns. Pyrex glass covers the etched silicon wafer to visualize the flow behavior inside the PHPs. The PHPs have rectangular channels with a hydraulic diameter of 923 mu m. Ethanol and R-134a are used as the working fluids. Based on the experimental data, a criterion of a diameter ratio (D-h/D-h,D- (crit)) is proposed to quantify the effect of a flow behavior on the thermal performance of two types of PHPs. A CEPHP performs better than a CLPHP when D-h/D-h,D- crit < 0.5. In the case of PHPs filled with ethanol (D-h/D-h,D- crit = 0.29), the CEPHP shows 33% lower thermal resistance and 20% higher maximum allowable heat flux than the CLPHP. In this case, an oscillation mode is observed inside both types of PHPs. In contrast, a CLPHP performs better than a CEPHP when D-h/D-h,D- crit > 0.5. In the case of PHPs filled with R-134a (D-h/D-h,D- crit = 0.69), the CLPHP shows 72% lower thermal resistance and 117% higher maximum allowable heat flux than the CEPHP. In this case, a circulation mode is observed inside the CLPHP, while an oscillation mode is observed inside the CEPHP. To explain why one type of PHP performs better than the other, an average volumetric fraction is introduced. An average volumetric fraction in the condenser section is experimentally shown to be a major contributor to the thermal performance; an increase of the average volumetric fraction in the condenser section leads to an improvement of the thermal performance due to enhanced latent heat transfer in the condenser section.