If you're using a parametric EQ, don't have a ton of different bands next to each other, instead just make one band with a lower Q value. A bunch of bands with thinner Q values will lead to the EQ looking wavy or wiggly with bumps in the frequency response, but one band with a higher Q value will lead to a smoother slope. For instance, I have a single band at 10Hz with a Q value of 0.3, this leads to the smooth bass curve from 10-100Hz that you see in this image (that curve is the EQ compensation, not the final frequency response). This example is only 5 bands (and without them that green line would be perfectly flat).
The Q value is how sharp or wide you want the band to be, so you see the small dip at 6.5k is a higher Q value (3) leading to a sharper dip. Compare this to the bass boost (basically compensating for the bass rolloff of the HD800) which has a Q value of 0.3, and is much wider.
My advice would be to get some kind of way to visualize what changes you are making to the frequency response, it makes working with the EQ so much easier (I'm using Peace, which is for Equalizer APO). Also, look at your sennheiser graph, its frequencies are limited, so I would supplement it with Tyll's measurements which go down to 10Hz.
If you want, I can give you my values to give you a jumping off point.