Automated negotiation has been of particular interest due to the relevant role that negotiation plays among trading agents. This paper presents two types of agent architecture: Case-Based and Fuzzy, tomodel an agent negotiation strategy. At each step of the negotiation process these architectures fix the weighted combination of tactics to employ and the parameter values related to these tactics. When an agent is provided with a Case-Based architecture, it uses previous knowledge and information of the environment state to change its negotiation behaviour. On the other hand when provided with a Fuzzy architecture it employs a set of fuzzy rules to determine the values of the parameters of the negotiation model. In this paper we propose an evolutionary approach, applying genetic algorithms over populations of agents provided with the same architecture, to determine which negotiation strategy is more successful.