This paper describes advanced techniques that can be used to select the best cotton blend, best in terms of quality and price. The first step in this is to model the spinning process, i.e. to describe the relationships between cotton blend and process conditions, and process behaviour and yarn properties. Two techniques are used for this, namely neural networks and learning classifier systems. The second step is to find the cotton blend and process conditions that allow to spin the required yarn at the best price. For this genetic algorithms have been applied. The research shows that the mentioned techniques have excellent potential to help the spinner to select the optimal cotton blend.