As rapid growth of the automotive exhaust emissions and fossil fuel consumption, biofuels become one of the reliable alternative fuels for automotive engines. Generally, the biofuels are blended with conventional fuels in different specific ratios. In this project, three different kinds of blending strategy are reviewed and the dual-injection strategy is used to modify a spark ignition engine for experimental testing. Moreover, the performance of automotive engines is sensitively affected by abrupt changes in fuel blend ratios, fuel variation and various environmental conditions. And among all the engine performance parameters, air-fuel ratio (AFR) is the most significant one as it relates closely to engine emissions, power output and fuel efficiency. Thus, an intelligent adaptive engine control system is necessary for maintaining the engine AFR under different operating conditions. Then several control methods are reviewed in this thesis. Obviously, traditional techniques are not adequately robust and too complex for the aim of adaptive control. Considering the limitations of traditional controllers, machine learning method is selected to design the controller. Then an intelligent adaptive controller based on backpropagation (BP) is proposed. In order to evaluate the performance of the proposed controller, a comparison between proposed control algorithm and typical PID method is carried out. In addition, in order to verify the effectiveness of the designed engine system, the engine in-cylinder pressure and AFR tests are conducted. Moreover, the structural reliability of the tailor-made engine test rig is originally analyzed through finite element method.