Gaussian boson sampling (GBS) is emerging as a powerful tool in solving complex graph problems, pivotal in advancing drug discovery. Traditional limitations in quantum computing, particularly in hardware scale and programmability, have hindered its practical applications. Addressing these challenges, we introduce a novel time-bin-encoded GBS photonic quantum processor. Characterized by its universality, programmability, and software scalability, our processor distinguishes itself with adjustable squeezing parameters and the capacity to execute arbitrary unitary operations. Utilizing this advanced architecture, we demonstrated clique finding on a 32-node graph with approximately double the success probability of classical approaches. Further, we employed our GBS processor in a ground-breaking quantum drug discovery platform, enabling efficient molecular docking and RNA-folding predictions. This work not only showcases the feasibility of GBS in complex computational tasks but also marks a significant stride towards its application in practical scenarios, particularly in drug discovery and molecular biology.