In view of the high adiabatic temperature rise and rapid hydration heat release rate of pumped concrete, a hydration temperature rising rate model based on hydration degree (hydration degree model) was constructed combined with the adiabatic temperature rise tests. Parameter inversion of hydration degree model, hyperbolic model, single exponential model and compound exponential model was performed and compared by the ABC (artificial bee colony) algorithm. The hydration degree model was applied to the fitting of five sets of adiabatic temperature rise test results, and the four temperature rise models were performed a feedback simulation analysis on the temperature field of a concrete test block in a pump gate project in Shanghai. The simulated results show good agreement with the experimental data. Hydration degree model can fully describe the characteristics of the hydration heat release in different stages, and the parameters are fitted in steps, which can greatly avoid the over-fitting problem when too many parameters exist. The model can provide more accurate temperature simulation results for intelligent temperature control during the construction of pumped concrete structures.