The aim of this work is to create a tool that can help in making difficult investment decisions about the sequence and commissioning periods of new hydrocarbons fields, maximizing expected monetary value and considering infrastructure constraints across a group of assets in terms geological uncertainties. The approach is to use probabilistic calculations and the modern portfolio theory, adapted to the conditions of the sphere of petroleum exploration and production. The goal of optimization is finding a strategy for new assets development that would have the maximum profit with minimal economic or technological risk. Measures of spread of a random variable (a.g. NPV) around its expected value are used to measure an economic risk. Also, we propose a method of technological risk assessment, reflecting the probability of hydrocarbon production in the required range. To obtain a probability distribution of the objective function according to uncertainties in input data the stochastic Monte-Carlo simulation and "Latin hypercube" technique were used. Due to high dimensionality of the problem, heuristic optimization methods have been used to solve the optimization task. A proxy model, based on the change in recoverable reserves volumes, starting production rates and decline rates of hydrocarbon production for newly drilled wells is proposed in this paper. The model combined with the calculation of economic measures. The combination of simulation models with an optimization algorithm allows to analyze a set of mathematical equations, containing a number of decision variables (e.g., years of commissioning periods and/or the paces of drilling wells) and to determine their optimum values, leading to a solution with the desired ratios of the expected economic effect and risk value. As a result, the numerical experiment revealed that application of the portfolio optimization methods allows to find new non-obvious solutions. The result of portfolio analysis is not a final product of project works, but, however, may be a basis for explanation of optimal development scenarios for a group of assets and can help in finding non-obvious, but more effective investment and project decisions. A comparison of four heuristic algorithms in the single-criterion optimization problem was conducted. Conclusions about the effectiveness of the algorithms in terms of time and the fitness function values were made. The developed model and methods were used in task of searching the optimal development strategy for a group of real hydrocarbons fields.