The thesis focuses on valuation and hedging problems when the market is incomplete. The Örst essay considers the quadratic hedging strategy. We propose a generalized quadratic hedging strategy which can balance a short-term risk (additional cost) with a long-term risk (hedging errors). The traditional quadratic hedging strategies, i.e. self-Önancing strategy and risk-minimization strategy, can be seen as special cases of the generalized quadratic hedging strategy. This is applied to the insurance derivatives market. The second essay compares parametric and nonparametric measure-changing techniques. The essay discusses three pricing approaches: pricing via Esscher measure, via calibration and via nonparametric risk-neutral density; and empirically compares the performance of the three approaches in the metal futures markets. The last essay establishes the concept of stochastic volatility of volatility and proposes several estimation methods.