This document discusses using fuzzy logic for short term load forecasting. It begins by explaining the importance of load forecasting for power systems and defines short term load forecasting as predicting load from 1 hour to 1 week ahead. It then describes using time, temperature, and previous day's load as inputs to a fuzzy logic system to generate load forecasts. Triangular and trapezoidal membership functions are used to fuzzify the inputs. Fuzzy rules are developed and the MATLAB Simulink software is used to simulate the system. Load data from a 132kV substation is used. The results show the fuzzy logic approach provides good short term load forecasting performance by handling the ambiguity in load data.