It depends on what you mean by "measure"
The numbers I give in this answer are inexact. The purpose of this answer is to help you clarify what you mean to ask, so that you can ask a more targeted question of experts who might be able to answer it (like on physics stackexchange).
Unaided Direct Observation
The coarsest scale we can consider is a unit of distance that can be observed by direct observation, where we can count and enumerate things that have differed by that value. This is probably around the 0.1-millimeter level (10^-4). Below this level, we probably could tell that an object of size X and an object of size 2X are different, but we wouldn't be able to quantify how much different they were.
Unaided Indirect Observation
The next scale up would be the scale of what we can prove to exist using our five senses, even if we can't directly observe individual changes on this scale.
For example, with our sense of touch, we can feel differences in texture caused by even very small differences in size (for example, threadcounts in high-quality bedsheets versus medium-quality ones). We can't measure how much smaller the threads are, but we can tell that the texture is different on a scale we cannot observe. Thus, we can prove the existence of a scale smaller than our observations.
The limit here is probably not much smaller than the size of a human cell, or tens of micrometers (10^-5 meters).
Tool-assisted Direct Observation
The best tools I'm aware of right now are electron microscopes. These can get us down to 0.1 nanometers (10^-10 meters).
Tool-assisted Indirect Observation
By taking many direct measurements and constructing probability models, we can measure the sizes of different kinds of atoms. This is an indirect observation, because even with our best tools, we cannot observe or count single atoms, but we can tell that (for example), a sample containing mostly-pure uranium is made-up primarily of structures with nuclei of 12 femtometers (1.2m * 10^-14). Our precision with these sorts of indirect measurements goes to about the tenth of a femtometer (10^-16).
Implications of our Models
Physical models (which predicted the existence of nuclear power) assume that these nuclei are made of subatomic particles (neutrons, protons, and electrons), and can make some predictions about how many of each there must be in the nucleus of any given element. Since we can already estimate the size of the nuclei, we can then make some estimates about how large these subatomic particles are. We can prove that electrons are smaller than an attometer (10^-18), but we don't know how much smaller, yet.
I consider this measurement to be the practical limit of our technology. With better technology, we might be able to measure the radius of an electron - or at least reduce the maximum size an electron could be.
Limits of Our Models
Based on our models, we can conceptualize what the smallest observable change might be - that is to say if our theories are correct, then there is no possible technology that will allow us to observe changes smaller than 10^-35m. Even if changes smaller than this are possible, then any change smaller than this amount would be literally inconsequential - there would be no effect on the rest of the system.
Of course, it's possible that our models are incomplete, and there is yet another layer of physics which can account for consequences on a smaller scale than this.