Evidence has identified various temporal eating patterns (TEPs), defined as the timing, frequency and sequencing of eating occasions (EOs) throughout the day (1, 2), are associated with diet quality and obesity prevalence, although findings are inconsistent (1-4). These inconsistencies may result from differences in analytic methods and the input variables used to derive TEPs; however, studies directly comparing these methods are rare. This study aimed to examine and compare two commonly used methods, latent class analysis (LCA) and modified dynamic time warping (MDTW) based cluster analysis for deriving TEPs and their association with diet quality and obesity. This cross-sectional study analysed data from 672 adults aged 18-65 years in Victoria, Australia, who completed seven-day food diary through ‘FoodNow’ smartphone app as part of the ‘EveryDayLife’ survey (2017-2020). The presence or absence of EOs at each hour of the day were used as input variables for LCA, while average hourly energy intake was used for MDTW-based clustering. Diet quality was assessed using the dietary guideline index (DGI), and body mass index (BMI) was calculated from self-reported height and weight. The LCA and MDTW-based clustering methods were compared through visualization of patterns, membership overlap, kappa statistics, and variance (adjusted R²) for explaining diet quality and BMI using linear regression. Both methods identified three TEPs, with Class 1 and Cluster 1, the largest groups, showing peaks in conditional probabilities of EOs and energy intake at 7:00 to 9:00, 12:00, and 18:00 to 19:00, reflecting conventional Australian mealtimes. Class 2 and Cluster 2 were characterised by delayed lunch at 13:00 and later evening peaks after 19:00. Class 3 and Cluster 3, the smallest group, had modest evenly spaced EOs and energy intake concentrated during earlier in the day (7:00 to 12:00). These TEPs showed moderate membership overlap (66.8% for Class 1 and Cluster 1, 56.2% for Class 2 and Cluster 2, and 73.1% for Class 3 and Cluster 3) with fair agreement (κ = 0.37, p < 0.001). Class 1 had more favourable diet quality than Class 2 (β = 4.8 ± 1.2, p < 0.001), which had the lowest diet quality among the three classes. Cluster 1 had more favourable diet quality than Cluster 2 (β = 2.6 ± 1.3, p = 0.044). No significant associations were observed between TEPs and BMI. LCA explained slightly more variance in diet quality (6% vs 4%) and similar variance in BMI (~11%) with MDTW-based clustering. In conclusion, both methods identified comparable TEPs, yet were not interchangeable. LCA was preferable when diet quality was the primary outcome, whereas MDTW-based clustering was more effective for identifying temporal patterns of energy intake, which could be advantageous when examining health outcomes. Researchers should choose methods aligned with their specific objectives.