Abstract: Estrus detection can greatly contribute to the reproductive performance and the economic profitability of modern pig production. An efficient and accurate estrus detection is often required to optimize the breeding timing for the small number of non-productive days, while the large number of piglets born per sow per year. However, the conventional detection can depend heavily on manual observation and the use of teaser boars, which are time-consuming, labor-intensive, and susceptible to subjectivity and biosecurity risks caused by direct contact between animals. Therefore, the innovative, automated, and non-invasive estrus detection technologies are critical to meet the growing demands of large-scale intelligent livestock farming. In this study, a comprehensive review was presented on the physiological and behavioral indicators of estrus in sows. Some indicators were highlighted, including the hormone levels, vulvar appearance, and behavioral patterns, such as standing reflex, restlessness, and mounting behavior. The biological basis was then formed to develop the detection systems. Some influencing factors were determined, such as the parity, breed, and environmental conditions. Three categories of estrus detection technologies were compared: the conventional manual observation, biological detection, and intelligent monitoring technologies. Manual observation, such as the back pressure tests and boar exposure, was required for the skilled experts prone to errors in the large-scale production, due to the simple and cost-effective operations. In biological detection, the molecular markers were utilized to be derived from omics approaches, including transcriptomics, proteomics, and metabolomics, in order to identify the estrus-related changes in saliva, urine, and blood samples. Specific genes, proteins, and metabolites were identified to be differentially expressed during estrus. The promising targets were then offered for the biomarker-based detection kits. The accurate performance was obtained using biological detection. The high costs, specialized equipment, and technical expertise limited their practical application on farms. Furthermore, the intelligent monitoring technologies have emerged as a transformative approach for automated estrus detection in recent years. Advanced sensing and data analysis techniques were utilized, including machine vision, infrared thermography, sound recognition, and wearable sensors. Machine vision algorithms were used to analyze the posture and behavioral patterns in order to detect estrus onset with high accuracy. Infrared imaging was used to capture the surface temperature in the vulvar and body regions, thus correlating with physiological variations during estrus. In sound analysis, the vocalization frequency and tone provided additional confirmation of estrus behavior. These multimodal systems were achieved in the binary classification accuracies of up to 95%, indicating the real-time, continuous monitoring with minimal animal stress. Several challenges also remained, including the robust spatiotemporal calibration over different sensor types, the limited algorithm generalizability under breeds and farm environments, and high initial investment costs for intelligent systems. A multidisciplinary approach was proposed to integrate edge computing, sensor fusion, lightweight algorithms, and scalable deployment, according to the needs of the commercial farms. Future research should focus on the integration of phenotype and molecular data for more comprehensive estrus prediction, the biosensor-based rapid detection tools for field application, and the welfare-oriented technologies for minimal animal stress. Moreover, the standardized datasets and benchmarking protocols can be expected to support the next-generation detection. In conclusion, the sow estrus detection can be evolved from conventional subjective methods to automatic, data-driven solutions that combine the biological insights with cutting-edge sensor technologies. Some advancements can improve the farm profitability to animal welfare in sustainable pig farming. The precision livestock industry can also realize the promising potential of intelligent breeding.