<p>In clinical medicine field, the issue of estimating the optimal individualized treatment regimes based on available characteristic information has garnered increasing attention due to the patients’ heterogeneous response to a special treatment. It is well known that all parametric methods in the study of the estimation problems are simple and interpretable but potentially suffer from model misspecification, while non-parametric methods are flexible but the resulting regimes are complicated so may not be interpretable in practice. In this paper, we use a semiparametric model to study the estimation of the optimal individualized treatment regime as it not only maintain the flexibility and interpretability but also can mitigate misidentification risks. We estimate the regimes via Bernstein polynomial and establish the asymptotic properties of the estimators. Both sufficient simulation studies and a real data analysis demonstrate that the proposed method has a good performance.</p>