Did you use an internal standard? It could be a matrix interference. You could also make a matrix matched calibration, some call it standard addition. 

what about anthracene d10 as internal standard 

It might work, but I don't think anthracene will ionize very well in the MS... I would try the standard addition, I bet your recoveries come out much better. 

1. You must blend the cabbage very well prior to extraction

2. You must mix the acetonitrile and sample well before adding the salts.

3. Use deuterated terphenyl as your internal standard - it's in the method (AOAC).

4. Clean up your extract - avoid using GCB in your dSPE mix if you can.

5. Are you sure you're not seeing degradation of the DDE in the injector port? Check your inertness - see USEPA method 8270 if you don't know how.

We spike with the analyte post-blending and prior to acetonitrile addition. For DDE you should be getting 80-120% recovery for fortified analytical portions (matrix spikes).

thanks mark  i use En standard method do you recomend specific internal standard , also do you recommend specific sim ions i use (316 ,318)  and (241 ,195 ) for endosulfan , for GCMSD with MMI do you inject sample in acetonitrile directly or evaporate it and dissolve in another sample 

Have you used proper clean-up procedure? There may have problem of matrix effect, though cabbage is not a very difficult matrix. Try not to use GCB for clean-up from cabbage matrix. I got good peaks of both 2,4-DDE and 4,4-DDE in GCMS from cabbage. In SIM mode you can have quantifier m/z as 246 and qualifier m/z(s) as 248 and 318. Both ACN and EA as solvent can give good results.