Machine Learning (ML) cloud services, offered by leading providers such as Amazon, Google, and Microsoft, enable the integration of ML components into software systems without building them from scratch. However, the rapid adoption of ML services, coupled with the growing complexity of business requirements, has led to a rise in misuses by software developers, compromising the quality, maintainability, and evolution of ML service-based systems. Though prior research has explored patterns and antipatterns in service-based and ML-based systems separately, automatic detection of ML service misuses remains a challenge. In this paper, we propose MLmisFinder, an automatic approach to detect ML service misuses in software systems, aiming to identify instances of improper use of ML services to help developers properly integrate ML components in ML service-based systems. To do so, we propose a generic, extensible, and reusable metamodel that captures the data needed to detect misuses in ML service-based systems and employs it to develop rule-based detection algorithms for seven types of misuse. We validated our approach on 107 software systems collected from open-source GitHub repositories and compared our results with a state-of-the-art (SOTA) baseline. Our results show that MLmisFinder effectively detects ML service misuses, achieving an average precision of 96.7% and recall of 97%, outperforming the SOTA baseline. Moreover, MLmisFinder demonstrated high efficiency and scalability in detecting misuses in 817 ML service-based systems and revealed a widespread prevalence of such misuses, especially in data drift monitoring and schema validation.