For securing a high-quality plating process, one of the main challenges in electroplating is the dosing of electrolytes. The optimal dosing and the related demand for resources do not only influence the coating quality but can also have a high environmental and economic relevance. Currently, the condition monitoring approaches related to dosing are mostly model-based and are seldom real-time capable. Therefore, a concept for real-time data-based condition monitoring of electrolytes is proposed. The paper discusses the challenges of pre-processing and modeling time series data with machine learning algorithms and quality requirements and availability of data within the electroplating process. Moreover, the usage of neural networks for condition monitoring of time series data is presented and discussed in a case study with a focus on anomaly detection. With this example, the applicability of a data-based approach for dynamic prediction of electrolyte chemicals is presented and evaluated.