In this paper, we extend the analysis of Cochrane and Saa-Requejo (2000) to deriving good-deal bounds on asset prices when investors worry about model uncertainty and seek robust pricing decisions in incomplete markets. Under the assumption that asset prices are driven by geometric Brownian motion processes, we propose a framework that is meaningful and very natural for investors' decision problems involving uncertainty about the pricing models, and derive closed-form solutions for the pricing bounds of the European option. We investigate properties of the proposed pricing bounds and apply these bounds to value a European option whose underlying asset is a non-traded stock index. We find that, under certain circumstances of model uncertainty, the proposed pricing bounds can contain sufficient amounts of the actual option prices, which is in contrast with the empirical finding of the good-deal bounds proposed by Cochrane and Saa-Requejo (2000).