This is not correct. In our models in relativity, "space", and more generally "spacetime", is defined as a manifold (3-dimensional for space, 4-dimensional for spacetime), with particular properties. All of this is perfectly well-defined mathematically. Physically, "points" in the manifold (for spacetime) correspond to events--physical happenings, such as "lightning strikes location X at time T by observer O's clock". The mathematical properties of the manifold correspond to the physical fact that, as far as we can tell, the set of physical happenings is continuous--there is no "minimum separation" in space or time between physical happenings. (There are speculations in quantum gravity that this may not hold at the Planck scale, but that scale is twenty orders of magnitude smaller than the smallest scale we can access experimentally, so the model of spacetime as a continuous manifold works at all the scales we can actually experiment with.)
Once you have spacetime as a manifold, "space" can be defined as some particular submanifold of spacetime, picked out according to some criterion (such as being a surface of constant coordinate time in some coordinate chart).