Thrombus is an extremely dangerous factor in the human body that can block the blood vessel. Once thrombus happens in venous of lower limbs, local blood flow is impeded. This leads to varicosity and even pulmonary embolism. In recent years, venous thromboembolism has frequently occurred in a variety of people, and there is no effective treatment for patients with different venous structures. For the patients with venous isomer with venous valve insufficiency, we establish a coupled computational model to simulate the process of thrombolysis with multi-dose treatment schemes by considering the blood as non-Newtonian fluid. And build corresponding in vitro experimental platform for verifying the performance of the developed mathematical model. Then, the effects of different fluid models, valve structures and drug doses on thrombolysis are comprehensively studied through numerical and experimental observations. Comparing with the experimental results, the relative error of blood boosting index (BBI) obtained from non-Newtonian fluid model is 11% times smaller than Newtonian fluid. In addition, the BBI from venous isomer is 13 times stronger than normal patient while the valve displacement is 5 times smaller. As consequence, low eddy current and strong molecular diffusion near the thrombus in case of isomer promote thrombolysis rate up to 18%. Furthermore, the 80 uM dosage of thrombolytic drugs gets the maximum thrombus dissolution rate 18% while the scheme of 50 uM doses obtains a thrombolysis rate of 14% in case of venous isomer. Under the two administration schemes for isomer patients, the rates from experiments are around 19.1% and 14.9%, respectively. It suggests that the proposed computational model and the designed experiment platform can potentially help different patients with venous thromboembolism to carry out clinical medication prediction.