It is necessary to predict the milling performance of different metal materials. This work develops a predictive model based on theoretical similarity reasoning model to evaluate cutting energy consumption and surface roughness in cast irons with varying matrix compositions. Using compacted graphite iron (CGI) and nodular cast iron (NCI) as test materials, this model achieves accurate prediction of NCI’s cutting power and surface roughness through CGI experimental data. The research reveals key machining parameter effects: cutting power increases while cutting force decreases with higher cutting speeds for CGI and NCI. Our work successfully predicts machining performance for cast irons with different graphite morphologies and matrix compositions under identical cutting parameters, providing both theoretical innovation and practical guidance that significantly reduces process development costs.