I was just listening to Dan Carlins podcast (Wrath of the Khans part I) when he mentioned this in passing when talking about the Steppe tribes. Is this correct? It seems kind of unlikely, especially given the definition of marriage used in the rest of his descriptions.
IIRC Carlin never said this about Mongols, he was talking about Sarmatians.
Their women, so long as they are virgins, ride, shoot, throw the javelin while mounted, and fight with their enemies. They do not lay aside their virginity until they have killed three of their enemies, and they do not marry before they have performed the traditional sacred rites. A woman who takes to herself a husband no longer rides, unless she is compelled to do so by a general expedition
-Hippocrates
I'm not exactly sure where Carlin pulled that claim from. It's definitely not something mentioned anywhere I've read, and is a rather broad statement. Really, because "the Mongols" were really a union of many previously semi-disparate tribes, it'd be like saying "did Western European women have to kill in battle before getting married?". Or "did Indian women have to kill in battle...?". They're broad groupings that don't always reflect that prior to and somewhat after the "Mongol Invasions", these are just groups of people who vary in their traditional customs to sometimes significant degrees. For example, there are Mongol tribes that lived closer to northern China that actually lived a fairly sedentary, agriculture/farming based life.
Now, as for the marriages, what would be the custom and the standard at least among most of the Mongols who would then come to rule the Yuan dynasty in China (aside from the emperors themselves) is levirate marriage: widows marrying their deceased husband'so brother. This was mainly a case of preserving kin and clan structure with bloodlines, and was a very important mechanism in achieving this.