The purpose of this study was to retrospectively describe the longitudinal changes of training variables in adolescent biathletes based on performance level. Thirty biathletes (15 men and 15 women) were included in the study and categorized as either national level biathletes (NLB, n = 21) or national team biathletes (NTB, n = 9). Retrospective training data was collected from training diary covering the biathletes’ four years (Y1-Y4) as student-athletes at upper secondary school. Training data was divided into physical and shooting training variables. A linear mixed-effect model was used for comparing the difference of the performance group and year of upper secondary school on training characteristics. The NTB group achieved a greater annual training volume than the NLB group, especially during Y4 (594±71 h·y-1 vs 461±127 h·y-1, p < 0.001), through an increase in duration of each session and by completing more weekly training volume during the general phase (13.7±4.6 vs 10.0±4.9 h·w-1, p = 0.004). No difference was observed in relative training intensity distribution between the groups. The total number of shots fired was also greater for the NTB (9971±4716 vs 7355±2812 shots·y-1, p = 0.003). There was an equal frequency in illness and injury for both the NLB and NTB. Accordingly, the results of the present study describe longitudinal changes of biathlon training in adolescent biathletes that also may affect performance development.