The invention discloses a power distribution transformer time sequence-based demand coefficient analysis method of historical acquisition data. By analyzing time sequence-based demand coefficient historical data, an inherent law is mined, and a development trend in future is further forecasted. By the method, the dynamic characteristic of a demand coefficient with time change is fully considered, the demand coefficient forecast value of next month is acquired by employing a data mining algorithm, the load ratio estimation capability of a power distribution transformer is improved, and data support is provided for relevant operation of a power grid department.