Middle height reinforced concrete frames are widely used in construction in residual areas and they can cause catastrophic disaster if they cant withstand during the destructive earthquakes. Hence determining the status of these buildings after earthquake and detecting mechanism formation are essential for safety insurance in the urban areas. This paper aims to determine the failure and non-failure modes of the flexural RC frames according to the damage status of the beam and column joints. To achieve this goal, a 5-storey flexural RC frame is modeled via IDARC software and Nonlinear Dynamic Time History Analysis is performed through 60 seismic accelerations. Then the frame collapse and non-collapse arrays are constructed obtaining the results of dynamic analysis in both modes. Artificial Neural Network is used for the Classification of the obtained modes. The results show good agreement in every class and make it possible to introduce the simple weight factor for frame status identification.