Gas concentration and temperature distribution reconstruction is realized based on tunable diode laser absorption spectroscopy. A measurement system using wavelength multiplexing technology with four absorption transitions is designed to reduce the number of projections and enrich the gas absorption measurement data. According to the nonlinear system established, a hybrid algorithm based on genetic algorithm and simulated annealing algorithm is introduced to obtain the global optimization and improve the search efficiency. A model for H2O concentration and temperature distribution in combustion is assumed, and numerical simulation is utilized to reconstruct the images with the transitions within 1.3—1.5 μm range. Both concentration and temperature distribution reconstruction results are in agreement with the model. The influence of random errors in projections on reconstruction is also analyzed. The increased error in projection would lead to larger mean square error of temperature reconstruction results, but little affect the concentration reconstruction results.