A simple, consistent, and self-contained error propagation algorithm was developed using the uncertainty information related to pure component physical properties, binary interaction parameters, and thermodynamic model parameters combined with Monte Carlo simulation along with the Latin Hypercube Sampling (LHS) method. This algorithm is generally applicable to simulate the error propagation in process flow sheets of arbitrary complexity as long as the thermodynamic model parameters encode uncertainty information. In this work, two significant problems related to hydrocarbon processing are studied under the light of uncertainty analysis. First, the injection of a valuable liquid hydrocarbon into an existing natural gas pipeline for transportation was studied in order to find the optimum injection rate of liquid n-butane that can be safely added to the flowing gas without undesired condensation. The main factors considered in this calculation are the hydrocarbon dew point, the natural gas physical properties, and conformity to pipeline specifications. Second, uncertainties on Reid vapor pressure (RVP) calculations were taken into account for the calculation of optimal rate of liquid n-butane blending into gasoline. Gasoline blending is an important operation in refineries where gasoline must be produced with enough volatility for the proper operation of engines in cold climates.