Network slicing (NS) is one of the major capabilities of the 5G network architecture. It is a key feature that leverages Network Function Virtualization (NFV) and Software-Defined Networking (SDN) to create isolated and flexible logical networks on top of physical infrastructure. It enables operators to automatically manage and optimize physical resources to meet heterogeneous service requirements. Various approaches have been employed to address the security challenges in the NS environment, especially the utilization of Machine Learning (ML) that benefits from vast amounts of data. However, traditional ML models fail to adequately ensure the confidentiality and privacy of sensitive training data. In addition, training ML models can face performance issues since all data is trained in a single model. Furthermore, coordinating defensive capabilities across different slices can be highly challenging. To address these concerns, this master thesis proposes Intra-Slice Aggregated Defender (ISAD), a security approach based on Federated Learning (FL) within the intra-slice ecosystem. The purpose of the proposed model (ISAD) is to preserve data privacy during training, detect anomalies within intra-slice ecosystem, prevent attacks based on model certainty, and improve the performance of model training while evaluating the security risks associated with each threat.