This paper presents a decision-support approach for balancing machining lines that is designed for on-to-fly use by shop-floor planners. The method relies on a declarative representation of the line-balancing problem and is solved with constraint programming, which makes it possible to quickly regenerate feasible production plans when priorities, orders, or operating conditions change. The proposed framework consolidates heterogeneous production information into a single, consistent data model and can be integrated with computer-aided planning and manufacturing supervision systems. To illustrate industrial relevance, the approach is evaluated in a real setting involving machining tasks for components such as rolling bearing parts and cast-iron hubs. The computational study highlights the benefits of an open, extensible architecture that allows company-specific rules and technological limitations to be added without redesigning the core model. Overall, the results indicate that the proposed solution improves responsiveness and transparency in dispatching decisions and can deliver operational advantages compared with more rigid balancing procedures.