Aiming at the goal of building Happy River and realizing river and lake governance in the new era, an evaluation index system of Happy River, including 23 core indexes and 27 optional indexes, was established based on ERG (existence relevance growth) demand model. Using MATLAB software, taking Happy River grade index as the target, the nonlinear relationship between Happy River evaluation index value and Happy River grade index was established by using projection pursuit model, and the Happy River grade evaluation model based on improved particle swarm optimization projection pursuit evaluation (IPSO-PPE) was constructed to realize the quantitative evaluation of Happy River. Taking Gansu section of the main stream of the Yellow River as an example, the change trend of Happy River grade index from 2010 to 2018 was analyzed. The results show that the grade index of Happy River in Gansu section of the main stream of the Yellow River shows an upward trend as a whole, from grade E (unhappiness) in 2010 to grade B (higher happiness index) in 2018. The coincidence rate between the evaluation results and the results of GRA-TOPSIS method is 88.9%, which can better reflect the actual situation of Gansu section of the Yellow River. Soil erosion, flood disaster in the basin and the satisfaction degree of river ecological base flow are the main factors restricting the grade index of Happy River in Gansu section of the Yellow River.