Insulin exhibits several temporal patterns, such as the 10- to 15-min pulsatile (minutes), additional (hours), and basal (days) secretions, leading to selective insulin responses in vivo; however, the mechanisms by which different temporal patterns of insulin selectively regulate downstream molecules remain unknown. Revealing the mechanisms of selective regulation by temporal patterns of insulin is pivotal for understanding insulin actions in vivo. We examined selective regulation of the insulin-Akt pathway and its mechanisms in the liver under hyperinsulinemic-euglycemic clamp conditions. We obtained time series data of the insulin-Akt pathway molecules using different stimulation patterns and developed a mathematical model that could reproduce these data. We found that all temporal patterns of the blood insulin levels are encoded into the insulin receptor (IR), and downstream molecules selectively and simultaneously decode them via protein kinase B (Akt or PKB). Mathematical modeling revealed the mechanisms via differences in network structures, sensitivity, and time constants. Moreover, we simulated the type II diabetes mellitus (T2DM) condition using the model and found that abnormal blood insulin patterns might contribute to the pathogenesis and/or progression of T2DM. Given that almost all hormones exhibit distinct temporal patterns, temporal coding may be a general principle of system homeostasis by hormones.