We determined player-to-player distance, body-to-ball contact, and exercise intensity during three training modalities in various football populations. 213 participants were recruited, ranging from 9-year-old boys to young men and 11-year-old girls to middle-aged women. All groups were analysed with video-filming and GPS-based Polar Pro monitors during three types of football training for 20 min, i.e., COVID-19-modified training (CMT) with >2-metre player-to-player distance, small-sided games (SSG), and simulated match-play with normal rules (SMP), in randomised order. Time spent in a danger zone (1.5 m) per-percent-infected-player (DZ PPIP) ranged from 0.015 to 0.279% of playing time. DZ PPIP for SSG was higher ( P < 0.05 ) than CMT and SMP. The average number of contacts (within 1.5 m) with a potentially infected player ranged from 12 to 73 contacts/hour. SSG had more ( P < 0.05 ) contacts than CMT and SMP, with SMP having a higher ( P < 0.05 ) number of contacts than CMT. Time/contact ranged from 0.87 to 3.00 seconds for the groups. No player-to-player and body-to-ball touches were registered for CMT. Total player-to-player contacts were 264% higher ( P < 0.05 ) in SSG than SMP, ranging from 80 to 170 and 25 to 56 touches, respectively. In all groups, a greater total distance was covered during SMP compared to CMT (38–114%; P < 0.05 ). All groups performed more high-intensity running (33–54%; P < 0.05 ) and had higher heart rates during SMP compared to CMT. Different types of football training all appear to exert a minor COVID-19 infection risk; however, COVID-19-modified training may be safer than small-sided game training, but also match-play. In contrast, exercise intensity is lower during COVID-19-modified training than match-play.