You can analyse the proximate composition of dehydrated plant materials. Calories are given by carbohydrates, proteins and the fat of the substance. You can analyse the protein content by Kjeldhal method. Crude fibre, fat, ash and moisture by AOAC methods. From the percentages obtained for protein, crude fat and (carbohydrates - fiber) you can calculate the calorie contribution considering the caloric contribution by each component. It is advisable to analyse the acid detergent fiber also to get an idea about the feed index. Hope this will help you.
It is est to determine total calories by bomb calorimetry as Valdo suggested. It is the most direct and precise method for the energy analysis of a specimen. I also believe in the micro Kjeldahl method for total protein. You might also consider analysing the ash and moisture contents as Suraji mentioned. The specimen is usually dried prior to these measurements, so wet and dry weights will enable calculations for total 'x' content of the natural specimen. Good luck
As suggested by some of the scientist, for determination of gross energy content you need to use a bomb calorimeter. You can use adiabatic or isoperibolic calorimeter (this last is more use nowdays as it is quicker). You need to remove the water content of the material, preferably by liophilization and homogenize after liophilization with a proper mill up to particules of 1mm of diameter preferably. For N determination you can use Kjedahl or Dumas combustion methods. The combustion to obtaine total N requires usally small sample size and it is performed in a few minutes..
Alternative method is to grow plant hydroponically using various nutrient composition. After the plant grown up, potential insect pest were then given. By comparing the number of insect per time period in plants of each nutrient composition, the value of nutrient in host plant, for the insect, then can be estimated.