Introduction: Burns are part of the management of introduced grasses in Chiapas, Mexico, and this may derive in forest fires. Objectives: To determine fuel load, fire behavior, CO2 emissions, and to get fire-fighting security issues for jaragua (Hyparrhenia rufa Nees.) grasslands and savannas. Methods: An artificial jaragua grassland and an artificial jaragua savanna were studied at the California and Flores Magón communities, respectively, in La Sepultura Biosphere Reserve. Were measured pre and post-fire fuel loads. Six prescribed burns (three heading and three backing fires) were conducted in each, grassland and savanna, and were measured meteorological as well as fire behavior variables. Emissions were estimated multiplying the consumed fuel load by an emission constant. Results: In the grassland, were obtained the following averages: fuel load, 6.214 t/ha; residual load, 0.107 t/ha, and CO2 emission, 10.449 t CO2/ha. For the savanna, were recorded 14.119, 2.161 and 20.460 t CO2/ha, respectively, without differences for the pre and post-fire fuel loads between heading and backing fires. For the grassland, the heading fires reached 3.92 m, 1.83 m and 22.3 m/min for flame length, flame depth and fire propagation rate, while for backing fires such values were, respectively: 1.07, 0.23 and 0.67 m/min, with significant differences. For the savanna, heading fires yielded 5.89 m (flame length), 1.53 m (flame depth) and 45.5 m/min (propagation rate), while for backing fires that values were 2.21, 0.76 and 2.8 m/min, also with significant differences. Conclusions: Under the studied environmental conditions, particularly in the savanna, fire behavior is dangerous so a good prescription and more care must be taken for conducting controlled or prescribed burns. For forest fires, direct firefighting by the head of the fire must be avoided, for is too dangerous; instead it is recommended a firefighting by the back of the fire as well as waiting for a backing fire-advance of it after the fire reaches a ridge. ...