First we have to develop a synthetic self-sustained, self-propagating ecosystem on earth (or in the lab), before anything like that can be exported to other planets. We haven't been able to do that yet, not even in computer models. One reason is because we haven't yet fully figured each and every single component of a complete ecosystem.
For illustration's sake, here's one simple example (note in the end how complex it gets),
You want to create an ecosystem which has 10 types of plants, a community of herbivores which predates the plants, a community of carnivores that predates the herbivores, and a microbial population that decomposes all dead organisms and produces nutrients for the plants. A complete cycle.
If the herbivore predates all the plants before they can reproduce, the ecosystem will collapse. If the carnivore predates all the herbivores before they reproduce, both will go extinct.
If you want a functional microbial community for plant nutrients, then you have to isolate each microbe and their function. How will nitrates distribute in this chain? Phosphates? Carbon, molybdenum, iron, manganese, chlorine, etc.. all these elements and compounds have to be recycled in the chain and passed up the hierarchy and back down. [We haven't fully figured out yet which organism is responsible for what]
Then there is intraspecific and interspecific competition between the organisms. Is one organism hoarding all the resources? One minor fuck up like this and the whole system can collapse.
Everything has to be completely balanced. We haven't really figured out yet how to do that in the lab. The major challenge is the ecology and lack of complete understanding of interaction between species.