Background: Whilst intravascular endoscopy can be used to identify lesions and assess the deployment of endovascular devices, it requires temporary blockage of the local blood flow during observation, posing a serious risk of ischaemia. Objective: To aid the design of a novel flow-blockage-free intravascular endoscope, we explored changes in the haemodynamic behaviour of the flush flow with respect to the flow injection speed and the system design. Methods: We first constructed the computational models for three candidate endoscope designs (i.e., Model A, B, and C). Using each of the three endoscopes, flow patterns in the target vessels (straight, bent, and twisted) under three different sets of boundary conditions (i.e., injection speed of the flush flow and the background blood flowrate) were then resolved through use of computational fluid dynamics and in vitro flow experiments. The design of endoscope and its optimal operating condition were evaluated in terms of the volume fraction within the vascular segment of interest, as well as the percentage of high-volume-fraction area (PHVFA) corresponding to three cross-sectional planes distal to the microcatheter tip. Results: With a mild narrowing at the endoscope neck, Model B exhibited the highest PHVFA, irrespective of location of the cross-sectional plane, compared with Models A and C which, respectively, had no narrowing and a moderate narrowing. The greatest difference in the PHVFA between the three models was observed on the cross-sectional plane 2 mm distal to the tip of the microcatheter (Model B: 33% vs. Model A: 18%). The background blood flowrate was found to have a strong impact on the resulting volume fraction of the flush flow close to the vascular wall, with the greatest difference being 44% (Model A). Conclusion: We found that the haemodynamic performance of endoscope Model B outperformed that of Models A and C, as it generated a flush flow that occupied the largest volume within the vascular segment of interest, suggesting that the endoscope design with a diameter narrowing of 30% at the endoscope neck might yield images of a better quality.