Smart cities increasingly depend on heterogeneous sensor networks that monitortraffic, weather, and air quality. Although these data streams are essential for urbanplanning, environmental monitoring, and decision support, they are often siloed, published in fragmented formats, and lack semantic integration. This fragmentationrestricts interoperability, hinders cross-domain analysis, and prevents systematicreuse in knowledge graphs and machine learning (ML) pipelines. We propose a modular Smart City Core Ontology that integrates traffic, weather, and pollution data at lane and intersection granularity. The ontology reuses estab-lished Semantic Web standards including SOSA for observations, OWL-Time fortemporal modeling, GeoSPARQL for spatial grounding, and QUDT for unit semantics in order to provide a unified semantic layer. Domain-specific modules captureoperational details, while the shared core ensures interoperability across domainsand supports extensibility. The ontology is evaluated through competency questions and instantiated withopen datasets from Darmstadt, Germany. Results show that it supports bothdomain specific queries such as lane-level traffic counts and pollution thresholdchecks, and cross-domain analytics such as traffic–weather correlations. Scalability tests confirm that it can process hundreds of thousands of daily observationswhile preserving logical consistency. By harmonizing multimodal urban data, the ontology enables knowledge graphconstruction and ontology-driven reasoning, while also preparing datasets for temporal knowledge graph embeddings. This creates a foundation for advanced MLtasks including anomaly detection, forecasting, and decision-support in smart cityoperations