Detecting deception in real interrogations for criminal cases is critically important. Interrogation is composed of evidence- driven conversation that calls for a need for proper integration of context, where most prior works treat it as a sequence mod- eling task. In this work, we propose a context-constrained sen- tence modeling approach for deception detection. Specifically, we introduce the use of a global context label that is defined on multi-sentences, i.e., a context label is marked as deception if any of its sentences are deceptive. Then, by using a con- textual integrator that aggregates predictions on local sentences for context label prediction, we improve deception detection by jointly optimizing global and local labels. Our approach sig- nificantly outperforms other models and achieves 76.38% and 73.15% in Unweighted Average Recall (UAR) at the local and global levels, respectively. We also conducted two analyses to further demonstrate the effectiveness of our approach.