Maintaining and promoting mental health requires understanding how emotions fluctuate throughout the day (Widershowben et al., 2019). To capture these fluctuations, methods such as the Experience Sampling Method (ESM), which require multiple daily reports, are often used. However, ESM places a considerable burden on participants and is limited in capturing highly time-resolved emotional dynamics, leaving it unclear how emotions vary within a single day. Recently, emoji-based reporting via smartwatches is less burdensome than traditional ESM and enables the collection of large volumes of in-situ emotional data (Watanabe et al., 2025). The present study aimed to identify within-day patterns of emotional change using a smartwatch-based emoji reporting system. 59 participants aged 20 to 39 (M = 28.89, SD = 6.24; 29 men, 30 women) were asked to report their emotions 16 times per day for a week (responding every 30-45 minutes). Each report was submitted by selecting an emoji that best matched their current emotional state. Across the study period, 4,103 valid reports were collected out of 5,600 possible responses (73.3%). Dynamic Time Warping and cluster analyses revealed three distinct trajectories of emotional fluctuation. The first cluster (n = 19) showed higher positive affect around midday and in the post-work period. The second cluster (n = 9) exhibited a gradual increase in positive emotions as the day progressed, peaking in the evening. The third cluster (n = 22) maintained relatively stable, neutral emotions with minimal fluctuation throughout the day. These findings demonstrate that emotional fluctuations in daily life are not uniform but can be classified into distinct trajectories, each characterized by different periods of positivity or stability. This study provides foundational insights into everyday emotional dynamics and offers empirical evidence that can inform the design of time-sensitive strategies for mental health promotion.