The phytophagy, biology, and thermal requirements of Tarsonemus confusus Ewing on strawberry were determined, together with the damage caused to flowers and fruits at different population densities under controlled conditions. In addition, a deep learning tool for early mite diagnosis was developed, based on the detection of tarsonemid-associated damage using the YOLOv11 model developed by Ultralytics, and validated in commercial fields. Mites were collected from a commercial strawberry crop in Michoacán, Mexico (20°06'75.11"N, -102°38'78.35"W). A mite colony of T. confususwas maintained on strawberry flowers under laboratory conditions (25°C, 65% RH, and a 12:12 h L:D). Identification was performed using morphological characters based on the taxonomic keys of Lindquist (1986) and Nucifora & Vacante (2004). Biology, thermal requirements, and reproductive behaviour were determined at three temperature regimes (20, 25, and 28 °C) under a completely randomized design. Developmental times, lower developmental thresholds, and thermal constants were estimated using mathematical models and linear regression analyses. To assess damage, mite-free strawberry plants were inoculated with different population densities. Injuries to flowers and fruits were recorded, including blight- like symptoms on stamens and carpels, brown spotting on petals, deformities, and achene disorganization. Population density was quantified, and data were statistically analysed. Field sampling and image acquisition were also conducted using a Phantom 4 Pro V2.0 drone to validate mite presence in commercial fields. Finally, a deep-learning tool was developed using the YOLOv11Ultralytics model, based on both field and laboratory images for early damage detection. The dataset was integrated, labeled, and scaled to train a system capable of recognizing symptoms Tarsonemus confusus was found associated with strawberry flowers and fruits. Its development was significantly influenced by temperature, with the duration of the egg, larvae, nymph, and adult stages decreasing as temperature increased from 20 to 28 °C. Base temperatures ranged from 9.2 to 14.6 °C and stage-specific thermal constants were estimated, showing a linear relationship between temperature and developmental rate. Artificial infestations caused severe damage to flowers and fruits, including necrosis, deformities, and quality loss. Based on these symptoms, an image dataset was generated to train a deep learning model based on YOLOv11, with a recall of 0.85, mAP of 0.74, and precision of 0.73 aimed at early diagnosis under field conditions. Our study contributes new insights into the feeding behaviour habits of T. confusus and contributes to the development of novel technologies for early pest diagnosis.