Electric utility capacity expansion planning determines the type (i.e. nuclear, coal-fired, oil-fired, gasturbines, etc.), the size (measured in MW), and the online date of new power plants to meet the growing electrical demand. Recently, the importance of uncertainty handling has received much attention. This thesis presents an ``expected`` approach and a case study for an expansion planning under uncertainty. Uncertainties addressed in this thesis included the future electricity demand, future fuel prices, construction costs and random failures of power plants. Based on the observation that the shorter construction lead time adopts more flexibly to uncertainty than the longer one, and on the need to come up with an alternative approach that avoids the complexity of the conventional contingency approach to uncertainty handling, this thesis focuses only on the current-year decisions, but with expected future-year plans. It is shown that the underlying stochastic problem with uncertain data could be converted to a deterministic problem with their expected value under some regularity conditions. In this case, no major planning procedure or methodology changes are required to account for uncertainty, even in construction costs, operating costs, and demand. Unlike the conventional expectation of uncertain demand, this thesis introduces the concepts of a horizontal expected load duration curve to develop ``expected`` approach. It turns out in our expected approach that the introduction of the base load capacity had better be decided based on the low demand scenario as a safeguard to prevent costly expensive overinvestments. It is shown that a full-fledged model which explicitly includes the planning lead time and the contingent expansion plan could be formulated through a deterministic dynamic programming problem whose stages are merit order and whose states are the vector of cumulative capacity levels. The dynamic programming formulation is a new approach for uncer...