The SARS-CoV-2 outbreak led to an inevitable and drastic transition to online training systems. This study aimed to analyze the differences between live-streaming and pre-recorded training modalities in rhythmic gymnastics (RG) with coaches and gymnasts of different technical levels. A total of 238 coaches and 277 athletes affiliated with the Italian Gymnastics Federation (FGI) participated in the study. The data collection tool was a self-collected questionnaire structured in three sections: participant’s information, characteristics of live-streaming training, and characteristics of pre-recorded training. A 2 × 2 ANOVA was used for each numeric variable. A Pearson’s chi-squared test was used for each categorical variable. For the athletes, training frequency, motivation, and efficacy were significantly higher (p < 0.05) with live streaming (3.7 ± 1.5 day/week, 3.8 ± 0.9 score and 3.8 ± 0.8 score, respectively) than with a pre-recorded modality (2.2 ± 1.7 day/week, 3.1 ± 1.2 score and 3.7 ± 0.9 score, respectively), while for coaches, significant differences (p < 0.005) were found between the two modalities only for training frequency (live streaming, 3.6 ± 1.8 days/week vs. pre-recorded, 2.1 ± 1.7 days/week). The adherence (number of No:Yes) was significantly higher (p < 0.05) for the live-streaming modality than for the pre-recorded modality in gold athletes (1:74 vs. 14:61, respectively), silver athletes (12:190 vs. 28:174, respectively), and gold coaches (3:63 vs. 11:55, respectively), but it was not for silver coaches. Physical preparation was performed significantly (p < 0.005) more often (number of No:Yes) for live streaming than for the pre-recorded modality by gold athletes (9:66 vs. 34:41, respectively), silver athletes (25:177 vs. 77:125, respectively), gold coaches (8:58 vs. 37:29), and silver coaches (33:139 vs. 85:87, respectively). Free body technical preparation was performed significantly (p < 0.005) more often (number of No:Yes) for live streaming than for the pre-recorded modality by gold athletes (15:60 vs. 39:36, respectively), silver athletes (84:118 vs. 121:81, respectively), gold coaches (10:56 vs. 38:28), and silver coaches (60:112 vs. 105:67, respectively), while no differences were found for silver athletes’ and coaches’ technical preparations for apparatus training between the two modalities. In conclusion, live streaming had greater positive effects on RG training during home confinement. However, pre-recorded training could be more effective for some types of training, depending on the technical level of the athletes.