To cope with renewable energy consumption, renewable energy and coal fired flexible retrofit coordination is needed for planning and operation. A generator grid load energy storage flexible planning method is proposed containing renewable energy and coal fired flexible retrofit considering uncertainty scenarios generation. Existing work primarily focused on generator grid coordination, grid energy storage coordination, however, generator grid load energy storage coordinated planning is scarce under high proportional renewable energy integration. For drawbacks of current flexible planning issue，generative adversarial network method is firstly used to consider wind and solar uncertainty, analyze renewable energy uncertainty and coal fire flexible retrofit characteristic. On this basis, a long timescale and short timescale bi-level generator grid load energy storage stochastic planning model is built containing renewable energy and coal fire power units’ coordination. Scenarios accuracy indices and flexible deficiency indices are used to explore the relationship between uncertainty scenarios generation and planning results. Finally, the model is tested by the 301 nodes real system case, the effectiveness of the proposed method is verified. The results show, the proposed method is more effective in deep peak regulation at noon during summer and autumn, adaptable to reduce renewable energy curtailment and more economic in planning results.