This is a very interesting plot. I would, of course, expect the curve to be a non-linear fit (exponential or logarithmic).
[table="width: 500, class: grid]
[tr][td]missile1[/td][td]range (km)[/td][td]payload (kg)[/td][/tr]
[tr][td]Agni-1[/td][td]0700[/td][td]1000[/td][/tr]
[tr][td]Agni-1[/td][td]1200[/td][td]1000[/td][/tr]
[tr][td]Agni-2[/td][td]2000[/td][td]1000[/td][/tr]
[tr][td]Agni-2[/td][td]2500[/td][td]1000[/td][/tr]
[tr][td]Agni-3[/td][td]3000[/td][td]2500[/td][/tr]
[tr][td]Agni-3[/td][td]3500[/td][td]2500[/td][/tr]
[tr][td]Agni-3[/td][td]5000[/td][td]2000[/td][/tr]
[tr][td]Agni-4[/td][td]2500[/td][td]2000[/td][/tr]
[tr][td]Agni-4[/td][td]3700[/td][td]2500[/td][/tr]
[tr][td]Agni-5[/td][td]5000[/td][td]1500[/td][/tr]
[/table]
I used the above data, collected from various sources, and is probably noisy. This is what it looks like, and doesn't look like any curve can be fit. Of course, I need more data to make a better curve-fit.