The integration of Artificial Intelligence (AI) into mental health care has ushered in a paradigm shift in how emotional well-being is assessed, monitored, and treated. Among the various AI applications, sentiment and emotion analysis has emerged as a vital tool in extracting psychological insights from unstructured data sources such as clinical notes, therapy transcripts, social media interactions, and mobile health applications. Leveraging advanced models like BERT (Bidirectional Encoder Representations from Transformers) and GPT (Generative Pre-trained Transformer), alongside psycholinguistic tools such as LIWC (Linguistic Inquiry and Word Count) and VADER (Valence Aware Dictionary and sEntiment Reasoner), AI systems are now capable of understanding and interpreting nuanced emotional expressions in text and speech. This review paper presents a comprehensive synthesis of current research and methodologies related to AI-driven sentiment and emotion detection in the context of mental health. We explore classical and deep learning approaches, hybrid models, and multimodal frameworks applied to diverse datasets including clinical conversations, patient self-reports, and public online content. Real-world applications such as AI-powered chatbots, teletherapy platforms, and real-time monitoring tools are examined in detail. In addition, we discuss the ethical implications, including data privacy, algorithmic bias, and interpretability, which are critical for the safe deployment of AI systems in healthcare settings. The paper concludes with a set of recommendations for future research, emphasizing the need for multimodal integration, real-time analytics, and personalized mental health interventions. This work aims to inform researchers, clinicians, and developers about the current landscape and potential of AI in advancing mental health care.