Decision-making is a complex process in which our brain has to combine different sources of information, such as noisy sensory evidence and expected reward, in appropriate ways to obtain the outcome that satisfies the decision-maker. Despite various studies on perceptual decision-making and value-based decision making, it is still unclear how the brain combines sensory and reward information to make a complex decision. A prime candidate for mediating this process is the basal ganglia pathway. This pathway is known to make separate contributions to perceptual decisions based on the interpretation of uncertain sensory evidence and value-based decisions that select among outcome options. To begin to investigate what computations are performed by the brain, particularly in the basal ganglia, we trained monkeys to perform a reward-biased visual motion direction discrimination task and performed single-unit extracellular recordings in the caudate nucleus, the input station in the basal ganglia. Fitting the monkeys’ behaviors to a drift-diffusion model, we found that the monkeys used a rational heuristic to combine sensory and reward information. This heuristic is suboptimal but leads to good-enough outcomes. We also found that the monkeys’ reward biases were sensitive to the changes in the reward functions from session to session. This adaptive adjustment could be a possible reason underlying the individual variability in their decision strategies. By recording in the caudate nucleus, we found that it is involved in both the decision-formation and evaluation: before the monkey started accumulating sensory evidence, the caudate neurons represented the reward context that could be used to form a reward bias; during decision-formation, some caudate neurons jointly represented sensory evidence and reward information, which could facilitate the combining of sensory and reward information appropriately. After a decision is made, caudate nucleus represented both decision confidence and reward expectation, two evaluation-related quantities that influence the monkeys’ subsequent decision behaviors.