Fractal-like aggregate (FA-) drag has been previously calculated/correlated/reported, but “mobility diameter” information is not sufficient to make rational calculations of Brownian coagulation rates (for, say, population-balance modeling). Indeed, until now, only conjectures about gyration-radius scaling behavior have been used to predict FA–FA collision cross sections! But such “scaling relations” are untrustworthy even for FA momentum-, energy-, and mass-transfer purposes, and improved FA-collision rate constants (appearing as “kernels” in the coagulation balance integro-PDE) are overdue. Our premise is that FA collision rates in the free-molecule regime can be predicted using a gas-kinetic type formulation. If ( a) carrier gas mean free path and FA persistence length are much larger than any characteristic FA size, ( b) FA number density is low, ( c) FA velocity and position are uncorrelated, and ( d) there is a “hard-sphere” interaction between primary particles of different FAs, such a theory is developed/applied here. We introduce an effective collision diameter, 〈 d ij 〉, depending on the geometries of the two participating FAs. Quasi-MC calculations are reported for large ensembles of pairs of FAs, each computer-generated using a tunable cluster–cluster (CC)-algorithm. Our results differ from frequently used theoretical estimates based only on FA gyration (or mobility) radii and D f . They also confirm that, if the size disparity of the colliding FAs is large, 〈 d ij 〉 obtained by simply assigning individual diameters to each FA are significantly overestimated. Modified collision rate expressions for FA-coagulation modeling are suggested.