You can try cutemouse, but I think you'll find it has *more* acceleration than DOSBox's internal mouse driver. Have you tried changing the sensitivity= setting in dosbox.conf to something less than the default of 100?

Thanks for tip, I tried it, and you are right, the acceleration is slightly worse with cutemouse driver, I will play with the sensitivity settings, thanks for the tip ripsaw.

Anyway if anyone has mouse driver where is acceleration option disabled, or I can configure it to minimal setting, I would appreciate it 😀

Edit: see ripsaw's post below for information on disabling the built-in mouse driver. Note this is not recommended for general use.

With the built-in driver disabled, I was able to load a Microsoft DOS mouse driver, such as "Microsoft Mousedriver V7.04 for DOS" (like so at the commandline: mouse.com /P4). The setting /P4 may sufficiently lessen the acceleration. It could be compared with running the built-in mouse driver with a low or default sensitivity setting. Edit: as noted below, the native mouse already has a flat response.

This mouse.com version should support many of these parameters:
Setting Default Values Switches
Sensitivity
Horizontal 50 5-100 /Hn
Vertical 50 5-100 /Vn
Hor & Ver 50 5-100 /Sn

Acceleration 2 1,2,3,4 /Pn

Hardware Cursor
Support off on/off /Y

Some mouse drivers check to see if a driver is already installed, some don't. However, you don't need to modify DOSBox to disable the internal mouse driver, just point INT 33h at an IRET instruction, which can be done with the program here.

My idea of "acceleration" in mouse drivers is the pointer moving farther when you move the mouse faster. In that sense, DOSBox's internal mouse driver has very flat (linear) response to the mouse input coming from SDL, but acceleration from the host can be a factor.

Very interesting, thanks! I missed that thread which contained your post with workaround. 😀 I tried it out. I did scan some of your previous posts on the topic and found one where you explained the flat response of the dosbox mouse to user input.

I wondered, too, about the host factor and whether the OP tested the software on a different host. Past threads indicate that host mouse acceleration may influence the dosbox mouse (caused by both special mouse software along with the internal Windows mouse setting). Other threads indicate a possibility of using autolock=false, although that may not be a favorable solution in any case. I edited the above post so future readers are pointed to the best information.

I got problems with mouse cursor too. Moving cursor to the left works fine, but moving to the right stucks. If I move mouse in circles the cursor went to the left border. Looks like mouse acceleration blocks the smooth movement. I tried change sensitivity and alternative DOS mouse drivers without success.
Now I tried to change settings of my Windows mouse drivers. If I change "signalrate" to the lowest value (125 signals/second) then the cursor in dosbox moves a little bit smoother. Is there a connection between signalrate and right-move-stuck?

Using G500 Refresh with Logitech Gaming Software 8.58.183 under Windows 7 Home x64
DOSBox 0.74
Tried to play "Mad TV"

When moving the mouse slowly in Mad TV there is some "resistance" in moving the pointer to the right. This behavior is a problem with the game itself, not a problem with DOSBox or real mouse drivers. A user has suggested a modification to the game executable on the Compatibility List that supposedly eliminates the unusual mouse driver parameters, but I haven't tried it:
http://www.dosbox.com/comp_list.php?showID=173

Wow!
Looks like working. Cursor moves smooth! Many thanks.

EDIT: Played complete round. Mouse works fine and stable.