Carbon capture followed by utilization or storage enables carbon-intensive sectors to abate their CO2 emissions. However, complexity and nonlinearity of the capture processes hinder the incorporation of their first-principles models into analyses and optimizations of overall carbon management strategies. Accordingly, it is desirable to have a systematic method to develop a computationally less demanding surrogate model, which can replace the rigorous CO2 capture process model, for use in a high-level decision-making environment. For such purposes, the surrogate model should be able to provide multiple information needed to connect to subsequent processing steps under varying source stream conditions. This research addresses the development of surrogate models for CO2 capture processes that enjoy significantly lowered complexity while preserving the key information. A surrogate model can be constructed by fitting the input-output data generated by process simulation and optimization with the rigorous model. Following the proposed method, surrogate models for the amine-based CO2 capture processes with two representative types of amines, monoethanolamine (MEA) and piperazine (PZ), are constructed and validated. The constructed surrogate models predict the specific steam consumption rate and total equipment purchase cost based on the input information of desired capture rate and CO2 source stream condition. The predicted information is shown to agree well with the simulation result of the rigorous firstprinciples model. This surrogate modeling approach can be applied to compare different capture technologies in the context of analyzing and synthesizing a larger CCUS processing network.