The fit sequential mode which fits each atomic model after subtracting the density occupied by other atomic models appears to be working find in ChimeraX 1.4 in my tests.  So I think it has something to do with your map and models and their initial placement.

  Here is an test example I just tried.  I opened a dimer atomic model, made a 10A map from it, split the dimer into two monomer models, move one of the monomers so it collides with the other, then fit the two monomers.

open 7ej3

molmap #1 10

split #1

color #1.2 lightblue

move x 10 model #1.1 coord #1.1

fit #1.1-2 in #2 seq 2 res 10

It worked as expected pushing the monomers apart so they fit nicely without clashing.

  To debug what is going wrong in your specific case you can do the sequential fitting steps by hand.  Make a simulated map from your second monomer with the molmap command, subtract it from your map using volume subtract with the "minrms true" option so the simulated map is first scaled to match the experimental map, then fit the first monomer in that difference map.  Now make a molmap simulated map of the new first monomer position, subtract it from the original experimental map and fit the second monomer into the result.  This is exactly the process that the fitmap command with the sequential option is using.  Doing it step by step will probably reveal your problem.  A simpler approach could be to use "fit #2,3 in #1 seq 1 res 3.7" to see just the first step (fitting only model #2 after subtracting the density occupied by #3), judge whether the result makes sense, then use "fit #3,2 in #1 seq 1 res 3.7" to take one more step fitting only model #3 after subtracting the density occupied by #2.

Tom

Sequential fitting resulting from example commands on 7ej3