This study aimed to (1) compare “in-situ” monitored acceleration-speed (AS_in-situ) profile metrics from training/competition data in elite female soccer players to similar metrics from profiles developed from isolated maximal sprint efforts (AS_sprint) and;​ (2) compare the confidence interval (CI) and a Tukey boxplot (BP) outlier removal technique on the training/competition data to derive AS_in-situ profiles. Fifteen national team soccer players participated in a 4-week camp while wearing 10 Hz GNSS units. Towards the middle of the camp, 2 × 40 m isolated maximal sprints were performed. AS_in-situ profiles (theoretical maximum acceleration A₀ in m∙s−2 and speed S₀ in m∙s−1) were computed using the CI and BP techniques with training/competition data. The sprint data were modelled separately to construct horizontal force–velocity (FV) profiles, from which AS_sprint profiles were derived. Bland-Altman analysis was used to assess agreement between the CI- and BP-derived AS_in-situ profiles to the AS_sprint profiles, as well as regression analysis for systematic and proportional bias. Additionally, 1-way ANOVAs with Tukey posthoc compared the metrics between each method of analysis. Using the BP method, good agreement of the AS_in-situ with AS_sprint profile metrics A₀/S₀ was displayed, whereas good to moderate agreement was shown for the CI. The CI technique showed a proportional bias for A₀/S₀. Good to excellent intertrial reliability was demonstrated for isolated sprint metrics. Both BP and CI techniques provided comparable AS_in-situ profiles to AS_sprint profiles. This current research demonstrated that AS_in-situ profiling is applicable in elite women’s soccer and will have further application in many team sports.