Thromboembolism is a pathological condition in which thrombus fragments detach, circulate through the bloodstream, and obstruct vessels, leading to potentially life-threatening conditions. Thrombolysis remains a widely used clinical therapy for thromboembolism; however, its efficacy is influenced by multiple factors, such as the composition and microstructure of thrombi, and local hemodynamics. Existing thrombolysis studies often focus on a specific perspective, without adequate consideration of the intricate interplays among multiple processes that occur on different spatial-temporal scales. In addition, previous studies [1, 2] have reported different clot lysis patterns (e.g., face dissolution, uniform dissolution). However, the mechanisms underlying these patterns, as well as the relationship between dissolution pattern and the overall dissolution rate, have yet to be elucidated. This study integrates numerical simulation [2] with theoretical analysis including scaling analysis and perturbation methods to elucidate the governing mechanisms of clot lysis across multiple scales. Our results show that clot lysis pattern can be classified into face, transition and uniform regimes based on effective Peclet number and Damköhler number. Furthermore, analytical expressions are obtained for the thickness of the dissolution layer, and the average clot lysis rate and lysis time. Finally, a reduced-order model is proposed for uniform clot lysis, and optimal drug delivery strategies are explored for different dissolution regimes. This study provides theoretical insights into the mechanism of thrombolysis, which can be used to guide the optimization of thrombolytic drug design and therapeutic strategies.